BEGIN:VCALENDAR VERSION:2.0 PRODID:-//NCC Bulgaria - ECPv5.5.0.1//NONSGML v1.0//EN CALSCALE:GREGORIAN METHOD:PUBLISH X-WR-CALNAME:NCC Bulgaria X-ORIGINAL-URL:https://eurocc-bulgaria.bg X-WR-CALDESC:Events for NCC Bulgaria BEGIN:VTIMEZONE TZID:Europe/Helsinki BEGIN:DAYLIGHT TZOFFSETFROM:+0200 TZOFFSETTO:+0300 TZNAME:EEST DTSTART:20250330T010000 END:DAYLIGHT BEGIN:STANDARD TZOFFSETFROM:+0300 TZOFFSETTO:+0200 TZNAME:EET DTSTART:20251026T010000 END:STANDARD BEGIN:DAYLIGHT TZOFFSETFROM:+0200 TZOFFSETTO:+0300 TZNAME:EEST DTSTART:20260329T010000 END:DAYLIGHT BEGIN:STANDARD TZOFFSETFROM:+0300 TZOFFSETTO:+0200 TZNAME:EET DTSTART:20261025T010000 END:STANDARD END:VTIMEZONE BEGIN:VEVENT DTSTART;VALUE=DATE:20260712 DTEND;VALUE=DATE:20260718 DTSTAMP:20261023T101420 CREATED:20251219T091028Z LAST-MODIFIED:20251219T091028Z UID:6827-1783814400-1784332799@eurocc-bulgaria.bg SUMMARY:16th International High Performance Computing (HPC) Summer School\, 12-17 July\, 2026 in Perth\, Australia\, hosted by the Pawsey Supercomputing Research Centre. DESCRIPTION:NCC BULGARIA\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n NCC presenting the success story: NCC Bulgaria is the National Competence Centre for HPC and HPDA\, operating within the EuroCC3 network and EuroHPC ecosystem. Coordinated by the Institute of Information and Communication Technologies at the Bulgarian Academy of Sciences (IICT-BAS) in partnership with the Sofia University “St. Kliment Ohridski” (SU) and the University of National and World Economy (UNWE)\, it supports industry\, academia\, and public organizations through specialized training\, consulting\, technical expertise\, and access to national and European HPC resources. Scientific partners involved The Institute of Information and Communication Technologies (IICT-BAS) conducts fundamental and applied scientific research in the fields of high-performance computing (HPC)\, machine learning\, big data processing\, cybersecurity\, optimization\, financial modeling and mathematical modeling. The developed methods and technologies find application in medicine\, ecology\, industry\, transport\, financial analysis\, process optimization\, and digital transformation. Technical/scientific Challenge Constructing reliable investment portfolios requires balancing expected returns\, market volatility\, and potential losses under adverse market conditions. For a diversified universe of 32 global financial assets\, the large number of possible allocations makes constrained portfolio optimization computationally demanding. An additional challenge is that different risk measures\, such as the Sharpe ratio and Conditional Value-at-Risk (CVaR)\, may favour substantially different investment strategies. Portfolio decisions are also sensitive to the historical estimation period and the numerical sampling method. The research therefore investigates how Monte Carlo (MC) and randomized Quasi-Monte Carlo (QMC) algorithms can efficiently identify feasible portfolios while assessing numerical accuracy\, computational cost\, and sensitivity to market conditions. \n \n \n \n \n \n \n \n \n \n \n Solution Researchers at the Institute of Information and Communication Technologies of the Bulgarian Academy of Sciences (IICT-BAS) developed a reproducible Python-based framework for stochastic portfolio optimization using the HEMUS supercomputer. The methodology compares classical Monte Carlo sampling with Mersenne Twister and randomized Quasi-Monte Carlo methods based on scrambled Sobol and randomized Halton sequences. Candidate portfolios are generated under full-investment and long-only constraints\, with a maximum allocation of 30% per asset. We investigate two optimization objectives: maximizing the Sharpe ratio to improve risk-adjusted returns and minimizing historical CVaR to limit losses in the worst 5% of monthly observations. The analysis considers three historical estimation windows of 3\, 5\, and 10 years and five candidate budgets ranging from 128 to 2048. We evaluate each configuration through 50 independent randomized repetitions and compare it with deterministic numerical benchmarks using approximation errors\, portfolio characteristics\, and execution times. This computational framework enables systematic analysis of how sampling techniques\, risk objectives and historical periods influence portfolio selection. Scientific impact Methodology and Computational Framework The research presents a reproducible Python-based framework for stochastic portfolio optimization\, implemented on the HEMUS supercomputer at IICT-BAS. The study examines 32 global financial assets across 3-\, 5- and 10-year historical windows\, comparing Monte Carlo (with pseudorandom generator Mersenne Twister) and randomized Quasi-Monte Carlo (with scrambled Sobol and Halton) methods under Sharpe ratio maximization and CVaR minimization. Experiment 1 – Accuracy and Convergence of Stochastic Portfolio Search: Five sampling budgets (128–2048 candidate portfolios) and 50 independent repetitions per configuration were evaluated against deterministic numerical benchmarks. Increasing the number of candidates reduced approximation errors\, while differences between sampling methods remained modest\, with classical Monte Carlo generally requiring less execution time.Experiment 2 – Financial Risk – Return Comparison: The 5-year Sharpe-oriented portfolios achieved a mean annualized return of 9.15%\, while the 10-year CVaR-oriented portfolios recorded the lowest mean historical monthly CVaR of 4.40% and annualized volatility of 6.59%. These results highlight the trade-off between maximizing risk-adjusted returns and minimizing downside risk.Experiment 3 – Historical Window and Sampling Sensitivity: The 3-\, 5- and 10-year estimation windows produced different portfolio allocations and risk–return characteristics. The optimization objective and historical period generally had a greater influence on portfolio selection than the choice of sampling generator.Scientific Contribution The study demonstrates the potential of HPC-supported stochastic optimization for reproducible financial risk analysis and numerical benchmarking. The HEMUS supercomputer enabled extensive computational experiments\, providing a foundation for scalable portfolio optimization and future financial decision-support systems. Benefits:  Improved optimization accuracy: Larger sampling budgets reduce approximation errors relative to deterministic benchmarks.Risk-aware portfolio selection: Compares return-oriented Sharpe optimization with downside-risk-focused CVaR.Reliable numerical benchmarking: Evaluates accuracy\, stability and computational efficiency across MC/RQMC methods.Better understanding of market sensitivity: Reveals how historical estimation windows influence investment outcomes.HPC-enabled research: Supports repeated simulations and scalable numerical experimentation.Future financial applications: Provides a foundation for computational portfolio analysis and investment decision-support systems.\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Success story # Highlights: Keywords: HPC; Monte Carlo; Quasi-Monte Carlo; Portfolio Optimization; Sharpe Ratio; CVaR; Financial Risk.Technology: HEMUS Supercomputer; Python; Mersenne Twister; Scrambled Sobol and Halton; Stochastic Optimization.Industry sectors: Finance/Insurance; IT/HPC Systems\, Services & Software Providers.Scientific achievement: Development of a reproducible HPC-based framework for comparing stochastic portfolio optimization methods.Key results: Larger sampling budgets improved numerical accuracy\, while optimization objectives and historical windows significantly influenced portfolio risk–return characteristics.HPC contribution: HEMUS enabled repeated simulations\, algorithm benchmarking and computational performance analysis.\n \n \n \n \n Comparison of the 3 time horizons through Sharpe-ratio optimization. Among the sampled candidates\, the 5-year Sharpe solution provides the highest expected return and the strongest reward-to-risk trade-off\, at the cost of the highest volatility across the three horizons \n \n \n \n \n Figure 1. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 2. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 3. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Comparison of the 3 time horizons through CVaR (5%). At N=100\, Mean-CVaR shifts the selected portfolios toward lower-volatility regions. The 5-year solution achieves the highest expected return while slightly reducing volatility relative to the 3-year case\, whereas the 10-year solution is the most conservative in volatility terms. \n \n \n \n \n Figure 4. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 5. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 6. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 7. Sharpe vs CVaR: Cross-Horizon Comparison. Sharpe is higher at all three horizons. CVaR is lower at all three horizons. \n \n \n \n \n \n \n \n \n \n \n Contact: Silvi-Maria Gurova | IICT-BAS | silvimaria.gurova@iict.bas.bgTodor Gurov | IICT-BAS | todor.gurov@iict.bas.bgAleksandar Kirilov | IICT-BAS | alex.kirilov@iict.bas.bg URL:https://eurocc-bulgaria.bg/?tribe_events=16th-international-high-performance-computing-hpc-summer-school-12-17-july-2026-in-perth-australia-hosted-by-the-pawsey-supercomputing-research-centre ATTACH;FMTTYPE=image/jpeg:https://eurocc-bulgaria.bg/wp-content/uploads/2025/12/interHPCschool26.jpg END:VEVENT BEGIN:VEVENT DTSTART;VALUE=DATE:20260611 DTEND;VALUE=DATE:20260613 DTSTAMP:20261023T101420 CREATED:20260319T084138Z LAST-MODIFIED:20260319T084138Z UID:7125-1781136000-1781308799@eurocc-bulgaria.bg SUMMARY:Joint SeisSol & ExaHyPE User Workshop\, Training\, and Hackathon\, 11–12 June\, Munich\, Germany DESCRIPTION:NCC BULGARIA\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n NCC presenting the success story: NCC Bulgaria is the National Competence Centre for HPC and HPDA\, operating within the EuroCC3 network and EuroHPC ecosystem. Coordinated by the Institute of Information and Communication Technologies at the Bulgarian Academy of Sciences (IICT-BAS) in partnership with the Sofia University “St. Kliment Ohridski” (SU) and the University of National and World Economy (UNWE)\, it supports industry\, academia\, and public organizations through specialized training\, consulting\, technical expertise\, and access to national and European HPC resources. Scientific partners involved The Institute of Information and Communication Technologies (IICT-BAS) conducts fundamental and applied scientific research in the fields of high-performance computing (HPC)\, machine learning\, big data processing\, cybersecurity\, optimization\, financial modeling and mathematical modeling. The developed methods and technologies find application in medicine\, ecology\, industry\, transport\, financial analysis\, process optimization\, and digital transformation. Technical/scientific Challenge Constructing reliable investment portfolios requires balancing expected returns\, market volatility\, and potential losses under adverse market conditions. For a diversified universe of 32 global financial assets\, the large number of possible allocations makes constrained portfolio optimization computationally demanding. An additional challenge is that different risk measures\, such as the Sharpe ratio and Conditional Value-at-Risk (CVaR)\, may favour substantially different investment strategies. Portfolio decisions are also sensitive to the historical estimation period and the numerical sampling method. The research therefore investigates how Monte Carlo (MC) and randomized Quasi-Monte Carlo (QMC) algorithms can efficiently identify feasible portfolios while assessing numerical accuracy\, computational cost\, and sensitivity to market conditions. \n \n \n \n \n \n \n \n \n \n \n Solution Researchers at the Institute of Information and Communication Technologies of the Bulgarian Academy of Sciences (IICT-BAS) developed a reproducible Python-based framework for stochastic portfolio optimization using the HEMUS supercomputer. The methodology compares classical Monte Carlo sampling with Mersenne Twister and randomized Quasi-Monte Carlo methods based on scrambled Sobol and randomized Halton sequences. Candidate portfolios are generated under full-investment and long-only constraints\, with a maximum allocation of 30% per asset. We investigate two optimization objectives: maximizing the Sharpe ratio to improve risk-adjusted returns and minimizing historical CVaR to limit losses in the worst 5% of monthly observations. The analysis considers three historical estimation windows of 3\, 5\, and 10 years and five candidate budgets ranging from 128 to 2048. We evaluate each configuration through 50 independent randomized repetitions and compare it with deterministic numerical benchmarks using approximation errors\, portfolio characteristics\, and execution times. This computational framework enables systematic analysis of how sampling techniques\, risk objectives and historical periods influence portfolio selection. Scientific impact Methodology and Computational Framework The research presents a reproducible Python-based framework for stochastic portfolio optimization\, implemented on the HEMUS supercomputer at IICT-BAS. The study examines 32 global financial assets across 3-\, 5- and 10-year historical windows\, comparing Monte Carlo (with pseudorandom generator Mersenne Twister) and randomized Quasi-Monte Carlo (with scrambled Sobol and Halton) methods under Sharpe ratio maximization and CVaR minimization. Experiment 1 – Accuracy and Convergence of Stochastic Portfolio Search: Five sampling budgets (128–2048 candidate portfolios) and 50 independent repetitions per configuration were evaluated against deterministic numerical benchmarks. Increasing the number of candidates reduced approximation errors\, while differences between sampling methods remained modest\, with classical Monte Carlo generally requiring less execution time.Experiment 2 – Financial Risk – Return Comparison: The 5-year Sharpe-oriented portfolios achieved a mean annualized return of 9.15%\, while the 10-year CVaR-oriented portfolios recorded the lowest mean historical monthly CVaR of 4.40% and annualized volatility of 6.59%. These results highlight the trade-off between maximizing risk-adjusted returns and minimizing downside risk.Experiment 3 – Historical Window and Sampling Sensitivity: The 3-\, 5- and 10-year estimation windows produced different portfolio allocations and risk–return characteristics. The optimization objective and historical period generally had a greater influence on portfolio selection than the choice of sampling generator.Scientific Contribution The study demonstrates the potential of HPC-supported stochastic optimization for reproducible financial risk analysis and numerical benchmarking. The HEMUS supercomputer enabled extensive computational experiments\, providing a foundation for scalable portfolio optimization and future financial decision-support systems. Benefits:  Improved optimization accuracy: Larger sampling budgets reduce approximation errors relative to deterministic benchmarks.Risk-aware portfolio selection: Compares return-oriented Sharpe optimization with downside-risk-focused CVaR.Reliable numerical benchmarking: Evaluates accuracy\, stability and computational efficiency across MC/RQMC methods.Better understanding of market sensitivity: Reveals how historical estimation windows influence investment outcomes.HPC-enabled research: Supports repeated simulations and scalable numerical experimentation.Future financial applications: Provides a foundation for computational portfolio analysis and investment decision-support systems.\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Success story # Highlights: Keywords: HPC; Monte Carlo; Quasi-Monte Carlo; Portfolio Optimization; Sharpe Ratio; CVaR; Financial Risk.Technology: HEMUS Supercomputer; Python; Mersenne Twister; Scrambled Sobol and Halton; Stochastic Optimization.Industry sectors: Finance/Insurance; IT/HPC Systems\, Services & Software Providers.Scientific achievement: Development of a reproducible HPC-based framework for comparing stochastic portfolio optimization methods.Key results: Larger sampling budgets improved numerical accuracy\, while optimization objectives and historical windows significantly influenced portfolio risk–return characteristics.HPC contribution: HEMUS enabled repeated simulations\, algorithm benchmarking and computational performance analysis.\n \n \n \n \n Comparison of the 3 time horizons through Sharpe-ratio optimization. Among the sampled candidates\, the 5-year Sharpe solution provides the highest expected return and the strongest reward-to-risk trade-off\, at the cost of the highest volatility across the three horizons \n \n \n \n \n Figure 1. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 2. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 3. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Comparison of the 3 time horizons through CVaR (5%). At N=100\, Mean-CVaR shifts the selected portfolios toward lower-volatility regions. The 5-year solution achieves the highest expected return while slightly reducing volatility relative to the 3-year case\, whereas the 10-year solution is the most conservative in volatility terms. \n \n \n \n \n Figure 4. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 5. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 6. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 7. Sharpe vs CVaR: Cross-Horizon Comparison. Sharpe is higher at all three horizons. CVaR is lower at all three horizons. \n \n \n \n \n \n \n \n \n \n \n Contact: Silvi-Maria Gurova | IICT-BAS | silvimaria.gurova@iict.bas.bgTodor Gurov | IICT-BAS | todor.gurov@iict.bas.bgAleksandar Kirilov | IICT-BAS | alex.kirilov@iict.bas.bg URL:https://eurocc-bulgaria.bg/?tribe_events=joint-seissol-exahype-user-workshop-training-and-hackathon-11-12-june-munich-germany ATTACH;FMTTYPE=image/png:https://eurocc-bulgaria.bg/wp-content/uploads/2026/03/jointSesExaWorkshop26.png END:VEVENT BEGIN:VEVENT DTSTART;TZID=Europe/Helsinki:20260609T140000 DTEND;TZID=Europe/Helsinki:20260609T180000 DTSTAMP:20261023T101420 CREATED:20260601T090046Z LAST-MODIFIED:20260611T134133Z UID:7228-1781013600-1781028000@eurocc-bulgaria.bg SUMMARY:HPC seminar series (#9)\, 9 June 2026\, IICT DESCRIPTION:NCC BULGARIA\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n NCC presenting the success story: NCC Bulgaria is the National Competence Centre for HPC and HPDA\, operating within the EuroCC3 network and EuroHPC ecosystem. Coordinated by the Institute of Information and Communication Technologies at the Bulgarian Academy of Sciences (IICT-BAS) in partnership with the Sofia University “St. Kliment Ohridski” (SU) and the University of National and World Economy (UNWE)\, it supports industry\, academia\, and public organizations through specialized training\, consulting\, technical expertise\, and access to national and European HPC resources. Scientific partners involved The Institute of Information and Communication Technologies (IICT-BAS) conducts fundamental and applied scientific research in the fields of high-performance computing (HPC)\, machine learning\, big data processing\, cybersecurity\, optimization\, financial modeling and mathematical modeling. The developed methods and technologies find application in medicine\, ecology\, industry\, transport\, financial analysis\, process optimization\, and digital transformation. Technical/scientific Challenge Constructing reliable investment portfolios requires balancing expected returns\, market volatility\, and potential losses under adverse market conditions. For a diversified universe of 32 global financial assets\, the large number of possible allocations makes constrained portfolio optimization computationally demanding. An additional challenge is that different risk measures\, such as the Sharpe ratio and Conditional Value-at-Risk (CVaR)\, may favour substantially different investment strategies. Portfolio decisions are also sensitive to the historical estimation period and the numerical sampling method. The research therefore investigates how Monte Carlo (MC) and randomized Quasi-Monte Carlo (QMC) algorithms can efficiently identify feasible portfolios while assessing numerical accuracy\, computational cost\, and sensitivity to market conditions. \n \n \n \n \n \n \n \n \n \n \n Solution Researchers at the Institute of Information and Communication Technologies of the Bulgarian Academy of Sciences (IICT-BAS) developed a reproducible Python-based framework for stochastic portfolio optimization using the HEMUS supercomputer. The methodology compares classical Monte Carlo sampling with Mersenne Twister and randomized Quasi-Monte Carlo methods based on scrambled Sobol and randomized Halton sequences. Candidate portfolios are generated under full-investment and long-only constraints\, with a maximum allocation of 30% per asset. We investigate two optimization objectives: maximizing the Sharpe ratio to improve risk-adjusted returns and minimizing historical CVaR to limit losses in the worst 5% of monthly observations. The analysis considers three historical estimation windows of 3\, 5\, and 10 years and five candidate budgets ranging from 128 to 2048. We evaluate each configuration through 50 independent randomized repetitions and compare it with deterministic numerical benchmarks using approximation errors\, portfolio characteristics\, and execution times. This computational framework enables systematic analysis of how sampling techniques\, risk objectives and historical periods influence portfolio selection. Scientific impact Methodology and Computational Framework The research presents a reproducible Python-based framework for stochastic portfolio optimization\, implemented on the HEMUS supercomputer at IICT-BAS. The study examines 32 global financial assets across 3-\, 5- and 10-year historical windows\, comparing Monte Carlo (with pseudorandom generator Mersenne Twister) and randomized Quasi-Monte Carlo (with scrambled Sobol and Halton) methods under Sharpe ratio maximization and CVaR minimization. Experiment 1 – Accuracy and Convergence of Stochastic Portfolio Search: Five sampling budgets (128–2048 candidate portfolios) and 50 independent repetitions per configuration were evaluated against deterministic numerical benchmarks. Increasing the number of candidates reduced approximation errors\, while differences between sampling methods remained modest\, with classical Monte Carlo generally requiring less execution time.Experiment 2 – Financial Risk – Return Comparison: The 5-year Sharpe-oriented portfolios achieved a mean annualized return of 9.15%\, while the 10-year CVaR-oriented portfolios recorded the lowest mean historical monthly CVaR of 4.40% and annualized volatility of 6.59%. These results highlight the trade-off between maximizing risk-adjusted returns and minimizing downside risk.Experiment 3 – Historical Window and Sampling Sensitivity: The 3-\, 5- and 10-year estimation windows produced different portfolio allocations and risk–return characteristics. The optimization objective and historical period generally had a greater influence on portfolio selection than the choice of sampling generator.Scientific Contribution The study demonstrates the potential of HPC-supported stochastic optimization for reproducible financial risk analysis and numerical benchmarking. The HEMUS supercomputer enabled extensive computational experiments\, providing a foundation for scalable portfolio optimization and future financial decision-support systems. Benefits:  Improved optimization accuracy: Larger sampling budgets reduce approximation errors relative to deterministic benchmarks.Risk-aware portfolio selection: Compares return-oriented Sharpe optimization with downside-risk-focused CVaR.Reliable numerical benchmarking: Evaluates accuracy\, stability and computational efficiency across MC/RQMC methods.Better understanding of market sensitivity: Reveals how historical estimation windows influence investment outcomes.HPC-enabled research: Supports repeated simulations and scalable numerical experimentation.Future financial applications: Provides a foundation for computational portfolio analysis and investment decision-support systems.\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Success story # Highlights: Keywords: HPC; Monte Carlo; Quasi-Monte Carlo; Portfolio Optimization; Sharpe Ratio; CVaR; Financial Risk.Technology: HEMUS Supercomputer; Python; Mersenne Twister; Scrambled Sobol and Halton; Stochastic Optimization.Industry sectors: Finance/Insurance; IT/HPC Systems\, Services & Software Providers.Scientific achievement: Development of a reproducible HPC-based framework for comparing stochastic portfolio optimization methods.Key results: Larger sampling budgets improved numerical accuracy\, while optimization objectives and historical windows significantly influenced portfolio risk–return characteristics.HPC contribution: HEMUS enabled repeated simulations\, algorithm benchmarking and computational performance analysis.\n \n \n \n \n Comparison of the 3 time horizons through Sharpe-ratio optimization. Among the sampled candidates\, the 5-year Sharpe solution provides the highest expected return and the strongest reward-to-risk trade-off\, at the cost of the highest volatility across the three horizons \n \n \n \n \n Figure 1. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 2. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 3. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Comparison of the 3 time horizons through CVaR (5%). At N=100\, Mean-CVaR shifts the selected portfolios toward lower-volatility regions. The 5-year solution achieves the highest expected return while slightly reducing volatility relative to the 3-year case\, whereas the 10-year solution is the most conservative in volatility terms. \n \n \n \n \n Figure 4. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 5. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 6. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 7. Sharpe vs CVaR: Cross-Horizon Comparison. Sharpe is higher at all three horizons. CVaR is lower at all three horizons. \n \n \n \n \n \n \n \n \n \n \n Contact: Silvi-Maria Gurova | IICT-BAS | silvimaria.gurova@iict.bas.bgTodor Gurov | IICT-BAS | todor.gurov@iict.bas.bgAleksandar Kirilov | IICT-BAS | alex.kirilov@iict.bas.bg URL:https://eurocc-bulgaria.bg/?tribe_events=hpc-seminar-series-9-9-june-2026-iict END:VEVENT BEGIN:VEVENT DTSTART;TZID=Europe/Helsinki:20260514T140000 DTEND;TZID=Europe/Helsinki:20260514T150000 DTSTAMP:20261023T101420 CREATED:20260514T130040Z LAST-MODIFIED:20260611T133945Z UID:7224-1778767200-1778770800@eurocc-bulgaria.bg SUMMARY:HPC seminar series (#8)\, 14 May 2026\, IICT DESCRIPTION:NCC BULGARIA\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n NCC presenting the success story: NCC Bulgaria is the National Competence Centre for HPC and HPDA\, operating within the EuroCC3 network and EuroHPC ecosystem. Coordinated by the Institute of Information and Communication Technologies at the Bulgarian Academy of Sciences (IICT-BAS) in partnership with the Sofia University “St. Kliment Ohridski” (SU) and the University of National and World Economy (UNWE)\, it supports industry\, academia\, and public organizations through specialized training\, consulting\, technical expertise\, and access to national and European HPC resources. Scientific partners involved The Institute of Information and Communication Technologies (IICT-BAS) conducts fundamental and applied scientific research in the fields of high-performance computing (HPC)\, machine learning\, big data processing\, cybersecurity\, optimization\, financial modeling and mathematical modeling. The developed methods and technologies find application in medicine\, ecology\, industry\, transport\, financial analysis\, process optimization\, and digital transformation. Technical/scientific Challenge Constructing reliable investment portfolios requires balancing expected returns\, market volatility\, and potential losses under adverse market conditions. For a diversified universe of 32 global financial assets\, the large number of possible allocations makes constrained portfolio optimization computationally demanding. An additional challenge is that different risk measures\, such as the Sharpe ratio and Conditional Value-at-Risk (CVaR)\, may favour substantially different investment strategies. Portfolio decisions are also sensitive to the historical estimation period and the numerical sampling method. The research therefore investigates how Monte Carlo (MC) and randomized Quasi-Monte Carlo (QMC) algorithms can efficiently identify feasible portfolios while assessing numerical accuracy\, computational cost\, and sensitivity to market conditions. \n \n \n \n \n \n \n \n \n \n \n Solution Researchers at the Institute of Information and Communication Technologies of the Bulgarian Academy of Sciences (IICT-BAS) developed a reproducible Python-based framework for stochastic portfolio optimization using the HEMUS supercomputer. The methodology compares classical Monte Carlo sampling with Mersenne Twister and randomized Quasi-Monte Carlo methods based on scrambled Sobol and randomized Halton sequences. Candidate portfolios are generated under full-investment and long-only constraints\, with a maximum allocation of 30% per asset. We investigate two optimization objectives: maximizing the Sharpe ratio to improve risk-adjusted returns and minimizing historical CVaR to limit losses in the worst 5% of monthly observations. The analysis considers three historical estimation windows of 3\, 5\, and 10 years and five candidate budgets ranging from 128 to 2048. We evaluate each configuration through 50 independent randomized repetitions and compare it with deterministic numerical benchmarks using approximation errors\, portfolio characteristics\, and execution times. This computational framework enables systematic analysis of how sampling techniques\, risk objectives and historical periods influence portfolio selection. Scientific impact Methodology and Computational Framework The research presents a reproducible Python-based framework for stochastic portfolio optimization\, implemented on the HEMUS supercomputer at IICT-BAS. The study examines 32 global financial assets across 3-\, 5- and 10-year historical windows\, comparing Monte Carlo (with pseudorandom generator Mersenne Twister) and randomized Quasi-Monte Carlo (with scrambled Sobol and Halton) methods under Sharpe ratio maximization and CVaR minimization. Experiment 1 – Accuracy and Convergence of Stochastic Portfolio Search: Five sampling budgets (128–2048 candidate portfolios) and 50 independent repetitions per configuration were evaluated against deterministic numerical benchmarks. Increasing the number of candidates reduced approximation errors\, while differences between sampling methods remained modest\, with classical Monte Carlo generally requiring less execution time.Experiment 2 – Financial Risk – Return Comparison: The 5-year Sharpe-oriented portfolios achieved a mean annualized return of 9.15%\, while the 10-year CVaR-oriented portfolios recorded the lowest mean historical monthly CVaR of 4.40% and annualized volatility of 6.59%. These results highlight the trade-off between maximizing risk-adjusted returns and minimizing downside risk.Experiment 3 – Historical Window and Sampling Sensitivity: The 3-\, 5- and 10-year estimation windows produced different portfolio allocations and risk–return characteristics. The optimization objective and historical period generally had a greater influence on portfolio selection than the choice of sampling generator.Scientific Contribution The study demonstrates the potential of HPC-supported stochastic optimization for reproducible financial risk analysis and numerical benchmarking. The HEMUS supercomputer enabled extensive computational experiments\, providing a foundation for scalable portfolio optimization and future financial decision-support systems. Benefits:  Improved optimization accuracy: Larger sampling budgets reduce approximation errors relative to deterministic benchmarks.Risk-aware portfolio selection: Compares return-oriented Sharpe optimization with downside-risk-focused CVaR.Reliable numerical benchmarking: Evaluates accuracy\, stability and computational efficiency across MC/RQMC methods.Better understanding of market sensitivity: Reveals how historical estimation windows influence investment outcomes.HPC-enabled research: Supports repeated simulations and scalable numerical experimentation.Future financial applications: Provides a foundation for computational portfolio analysis and investment decision-support systems.\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Success story # Highlights: Keywords: HPC; Monte Carlo; Quasi-Monte Carlo; Portfolio Optimization; Sharpe Ratio; CVaR; Financial Risk.Technology: HEMUS Supercomputer; Python; Mersenne Twister; Scrambled Sobol and Halton; Stochastic Optimization.Industry sectors: Finance/Insurance; IT/HPC Systems\, Services & Software Providers.Scientific achievement: Development of a reproducible HPC-based framework for comparing stochastic portfolio optimization methods.Key results: Larger sampling budgets improved numerical accuracy\, while optimization objectives and historical windows significantly influenced portfolio risk–return characteristics.HPC contribution: HEMUS enabled repeated simulations\, algorithm benchmarking and computational performance analysis.\n \n \n \n \n Comparison of the 3 time horizons through Sharpe-ratio optimization. Among the sampled candidates\, the 5-year Sharpe solution provides the highest expected return and the strongest reward-to-risk trade-off\, at the cost of the highest volatility across the three horizons \n \n \n \n \n Figure 1. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 2. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 3. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Comparison of the 3 time horizons through CVaR (5%). At N=100\, Mean-CVaR shifts the selected portfolios toward lower-volatility regions. The 5-year solution achieves the highest expected return while slightly reducing volatility relative to the 3-year case\, whereas the 10-year solution is the most conservative in volatility terms. \n \n \n \n \n Figure 4. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 5. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 6. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 7. Sharpe vs CVaR: Cross-Horizon Comparison. Sharpe is higher at all three horizons. CVaR is lower at all three horizons. \n \n \n \n \n \n \n \n \n \n \n Contact: Silvi-Maria Gurova | IICT-BAS | silvimaria.gurova@iict.bas.bgTodor Gurov | IICT-BAS | todor.gurov@iict.bas.bgAleksandar Kirilov | IICT-BAS | alex.kirilov@iict.bas.bg URL:https://eurocc-bulgaria.bg/?tribe_events=hpc-seminar-series-8-14-may-2026-iict END:VEVENT BEGIN:VEVENT DTSTART;TZID=Europe/Helsinki:20260514T120000 DTEND;TZID=Europe/Helsinki:20260514T133000 DTSTAMP:20261023T101420 CREATED:20260316T135625Z LAST-MODIFIED:20260316T135625Z UID:7070-1778760000-1778765400@eurocc-bulgaria.bg SUMMARY:ENCCS Webinar Software Installation on HPC systems\, May 14th\, 12:00-13:30 (CET)\, NCC Sweden\, online. DESCRIPTION:NCC BULGARIA\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n NCC presenting the success story: NCC Bulgaria is the National Competence Centre for HPC and HPDA\, operating within the EuroCC3 network and EuroHPC ecosystem. Coordinated by the Institute of Information and Communication Technologies at the Bulgarian Academy of Sciences (IICT-BAS) in partnership with the Sofia University “St. Kliment Ohridski” (SU) and the University of National and World Economy (UNWE)\, it supports industry\, academia\, and public organizations through specialized training\, consulting\, technical expertise\, and access to national and European HPC resources. Scientific partners involved The Institute of Information and Communication Technologies (IICT-BAS) conducts fundamental and applied scientific research in the fields of high-performance computing (HPC)\, machine learning\, big data processing\, cybersecurity\, optimization\, financial modeling and mathematical modeling. The developed methods and technologies find application in medicine\, ecology\, industry\, transport\, financial analysis\, process optimization\, and digital transformation. Technical/scientific Challenge Constructing reliable investment portfolios requires balancing expected returns\, market volatility\, and potential losses under adverse market conditions. For a diversified universe of 32 global financial assets\, the large number of possible allocations makes constrained portfolio optimization computationally demanding. An additional challenge is that different risk measures\, such as the Sharpe ratio and Conditional Value-at-Risk (CVaR)\, may favour substantially different investment strategies. Portfolio decisions are also sensitive to the historical estimation period and the numerical sampling method. The research therefore investigates how Monte Carlo (MC) and randomized Quasi-Monte Carlo (QMC) algorithms can efficiently identify feasible portfolios while assessing numerical accuracy\, computational cost\, and sensitivity to market conditions. \n \n \n \n \n \n \n \n \n \n \n Solution Researchers at the Institute of Information and Communication Technologies of the Bulgarian Academy of Sciences (IICT-BAS) developed a reproducible Python-based framework for stochastic portfolio optimization using the HEMUS supercomputer. The methodology compares classical Monte Carlo sampling with Mersenne Twister and randomized Quasi-Monte Carlo methods based on scrambled Sobol and randomized Halton sequences. Candidate portfolios are generated under full-investment and long-only constraints\, with a maximum allocation of 30% per asset. We investigate two optimization objectives: maximizing the Sharpe ratio to improve risk-adjusted returns and minimizing historical CVaR to limit losses in the worst 5% of monthly observations. The analysis considers three historical estimation windows of 3\, 5\, and 10 years and five candidate budgets ranging from 128 to 2048. We evaluate each configuration through 50 independent randomized repetitions and compare it with deterministic numerical benchmarks using approximation errors\, portfolio characteristics\, and execution times. This computational framework enables systematic analysis of how sampling techniques\, risk objectives and historical periods influence portfolio selection. Scientific impact Methodology and Computational Framework The research presents a reproducible Python-based framework for stochastic portfolio optimization\, implemented on the HEMUS supercomputer at IICT-BAS. The study examines 32 global financial assets across 3-\, 5- and 10-year historical windows\, comparing Monte Carlo (with pseudorandom generator Mersenne Twister) and randomized Quasi-Monte Carlo (with scrambled Sobol and Halton) methods under Sharpe ratio maximization and CVaR minimization. Experiment 1 – Accuracy and Convergence of Stochastic Portfolio Search: Five sampling budgets (128–2048 candidate portfolios) and 50 independent repetitions per configuration were evaluated against deterministic numerical benchmarks. Increasing the number of candidates reduced approximation errors\, while differences between sampling methods remained modest\, with classical Monte Carlo generally requiring less execution time.Experiment 2 – Financial Risk – Return Comparison: The 5-year Sharpe-oriented portfolios achieved a mean annualized return of 9.15%\, while the 10-year CVaR-oriented portfolios recorded the lowest mean historical monthly CVaR of 4.40% and annualized volatility of 6.59%. These results highlight the trade-off between maximizing risk-adjusted returns and minimizing downside risk.Experiment 3 – Historical Window and Sampling Sensitivity: The 3-\, 5- and 10-year estimation windows produced different portfolio allocations and risk–return characteristics. The optimization objective and historical period generally had a greater influence on portfolio selection than the choice of sampling generator.Scientific Contribution The study demonstrates the potential of HPC-supported stochastic optimization for reproducible financial risk analysis and numerical benchmarking. The HEMUS supercomputer enabled extensive computational experiments\, providing a foundation for scalable portfolio optimization and future financial decision-support systems. Benefits:  Improved optimization accuracy: Larger sampling budgets reduce approximation errors relative to deterministic benchmarks.Risk-aware portfolio selection: Compares return-oriented Sharpe optimization with downside-risk-focused CVaR.Reliable numerical benchmarking: Evaluates accuracy\, stability and computational efficiency across MC/RQMC methods.Better understanding of market sensitivity: Reveals how historical estimation windows influence investment outcomes.HPC-enabled research: Supports repeated simulations and scalable numerical experimentation.Future financial applications: Provides a foundation for computational portfolio analysis and investment decision-support systems.\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Success story # Highlights: Keywords: HPC; Monte Carlo; Quasi-Monte Carlo; Portfolio Optimization; Sharpe Ratio; CVaR; Financial Risk.Technology: HEMUS Supercomputer; Python; Mersenne Twister; Scrambled Sobol and Halton; Stochastic Optimization.Industry sectors: Finance/Insurance; IT/HPC Systems\, Services & Software Providers.Scientific achievement: Development of a reproducible HPC-based framework for comparing stochastic portfolio optimization methods.Key results: Larger sampling budgets improved numerical accuracy\, while optimization objectives and historical windows significantly influenced portfolio risk–return characteristics.HPC contribution: HEMUS enabled repeated simulations\, algorithm benchmarking and computational performance analysis.\n \n \n \n \n Comparison of the 3 time horizons through Sharpe-ratio optimization. Among the sampled candidates\, the 5-year Sharpe solution provides the highest expected return and the strongest reward-to-risk trade-off\, at the cost of the highest volatility across the three horizons \n \n \n \n \n Figure 1. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 2. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 3. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Comparison of the 3 time horizons through CVaR (5%). At N=100\, Mean-CVaR shifts the selected portfolios toward lower-volatility regions. The 5-year solution achieves the highest expected return while slightly reducing volatility relative to the 3-year case\, whereas the 10-year solution is the most conservative in volatility terms. \n \n \n \n \n Figure 4. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 5. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 6. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 7. Sharpe vs CVaR: Cross-Horizon Comparison. Sharpe is higher at all three horizons. CVaR is lower at all three horizons. \n \n \n \n \n \n \n \n \n \n \n Contact: Silvi-Maria Gurova | IICT-BAS | silvimaria.gurova@iict.bas.bgTodor Gurov | IICT-BAS | todor.gurov@iict.bas.bgAleksandar Kirilov | IICT-BAS | alex.kirilov@iict.bas.bg URL:https://eurocc-bulgaria.bg/?tribe_events=enccs-webinar-software-installation-on-hpc-systems-may-14th-1200-1330-cet-ncc-sweden-online ATTACH;FMTTYPE=image/jpeg:https://eurocc-bulgaria.bg/wp-content/uploads/2026/03/EnccsWebinar26HPC.jpg END:VEVENT BEGIN:VEVENT DTSTART;VALUE=DATE:20260325 DTEND;VALUE=DATE:20260327 DTSTAMP:20261023T101420 CREATED:20260316T140015Z LAST-MODIFIED:20260316T140015Z UID:7075-1774396800-1774569599@eurocc-bulgaria.bg SUMMARY:Molecular Dynamics Simulation with GROMACS\, organized by NCC |Austria\, 25–26 Mar 2026\, ONLINE DESCRIPTION:NCC BULGARIA\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n NCC presenting the success story: NCC Bulgaria is the National Competence Centre for HPC and HPDA\, operating within the EuroCC3 network and EuroHPC ecosystem. Coordinated by the Institute of Information and Communication Technologies at the Bulgarian Academy of Sciences (IICT-BAS) in partnership with the Sofia University “St. Kliment Ohridski” (SU) and the University of National and World Economy (UNWE)\, it supports industry\, academia\, and public organizations through specialized training\, consulting\, technical expertise\, and access to national and European HPC resources. Scientific partners involved The Institute of Information and Communication Technologies (IICT-BAS) conducts fundamental and applied scientific research in the fields of high-performance computing (HPC)\, machine learning\, big data processing\, cybersecurity\, optimization\, financial modeling and mathematical modeling. The developed methods and technologies find application in medicine\, ecology\, industry\, transport\, financial analysis\, process optimization\, and digital transformation. Technical/scientific Challenge Constructing reliable investment portfolios requires balancing expected returns\, market volatility\, and potential losses under adverse market conditions. For a diversified universe of 32 global financial assets\, the large number of possible allocations makes constrained portfolio optimization computationally demanding. An additional challenge is that different risk measures\, such as the Sharpe ratio and Conditional Value-at-Risk (CVaR)\, may favour substantially different investment strategies. Portfolio decisions are also sensitive to the historical estimation period and the numerical sampling method. The research therefore investigates how Monte Carlo (MC) and randomized Quasi-Monte Carlo (QMC) algorithms can efficiently identify feasible portfolios while assessing numerical accuracy\, computational cost\, and sensitivity to market conditions. \n \n \n \n \n \n \n \n \n \n \n Solution Researchers at the Institute of Information and Communication Technologies of the Bulgarian Academy of Sciences (IICT-BAS) developed a reproducible Python-based framework for stochastic portfolio optimization using the HEMUS supercomputer. The methodology compares classical Monte Carlo sampling with Mersenne Twister and randomized Quasi-Monte Carlo methods based on scrambled Sobol and randomized Halton sequences. Candidate portfolios are generated under full-investment and long-only constraints\, with a maximum allocation of 30% per asset. We investigate two optimization objectives: maximizing the Sharpe ratio to improve risk-adjusted returns and minimizing historical CVaR to limit losses in the worst 5% of monthly observations. The analysis considers three historical estimation windows of 3\, 5\, and 10 years and five candidate budgets ranging from 128 to 2048. We evaluate each configuration through 50 independent randomized repetitions and compare it with deterministic numerical benchmarks using approximation errors\, portfolio characteristics\, and execution times. This computational framework enables systematic analysis of how sampling techniques\, risk objectives and historical periods influence portfolio selection. Scientific impact Methodology and Computational Framework The research presents a reproducible Python-based framework for stochastic portfolio optimization\, implemented on the HEMUS supercomputer at IICT-BAS. The study examines 32 global financial assets across 3-\, 5- and 10-year historical windows\, comparing Monte Carlo (with pseudorandom generator Mersenne Twister) and randomized Quasi-Monte Carlo (with scrambled Sobol and Halton) methods under Sharpe ratio maximization and CVaR minimization. Experiment 1 – Accuracy and Convergence of Stochastic Portfolio Search: Five sampling budgets (128–2048 candidate portfolios) and 50 independent repetitions per configuration were evaluated against deterministic numerical benchmarks. Increasing the number of candidates reduced approximation errors\, while differences between sampling methods remained modest\, with classical Monte Carlo generally requiring less execution time.Experiment 2 – Financial Risk – Return Comparison: The 5-year Sharpe-oriented portfolios achieved a mean annualized return of 9.15%\, while the 10-year CVaR-oriented portfolios recorded the lowest mean historical monthly CVaR of 4.40% and annualized volatility of 6.59%. These results highlight the trade-off between maximizing risk-adjusted returns and minimizing downside risk.Experiment 3 – Historical Window and Sampling Sensitivity: The 3-\, 5- and 10-year estimation windows produced different portfolio allocations and risk–return characteristics. The optimization objective and historical period generally had a greater influence on portfolio selection than the choice of sampling generator.Scientific Contribution The study demonstrates the potential of HPC-supported stochastic optimization for reproducible financial risk analysis and numerical benchmarking. The HEMUS supercomputer enabled extensive computational experiments\, providing a foundation for scalable portfolio optimization and future financial decision-support systems. Benefits:  Improved optimization accuracy: Larger sampling budgets reduce approximation errors relative to deterministic benchmarks.Risk-aware portfolio selection: Compares return-oriented Sharpe optimization with downside-risk-focused CVaR.Reliable numerical benchmarking: Evaluates accuracy\, stability and computational efficiency across MC/RQMC methods.Better understanding of market sensitivity: Reveals how historical estimation windows influence investment outcomes.HPC-enabled research: Supports repeated simulations and scalable numerical experimentation.Future financial applications: Provides a foundation for computational portfolio analysis and investment decision-support systems.\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Success story # Highlights: Keywords: HPC; Monte Carlo; Quasi-Monte Carlo; Portfolio Optimization; Sharpe Ratio; CVaR; Financial Risk.Technology: HEMUS Supercomputer; Python; Mersenne Twister; Scrambled Sobol and Halton; Stochastic Optimization.Industry sectors: Finance/Insurance; IT/HPC Systems\, Services & Software Providers.Scientific achievement: Development of a reproducible HPC-based framework for comparing stochastic portfolio optimization methods.Key results: Larger sampling budgets improved numerical accuracy\, while optimization objectives and historical windows significantly influenced portfolio risk–return characteristics.HPC contribution: HEMUS enabled repeated simulations\, algorithm benchmarking and computational performance analysis.\n \n \n \n \n Comparison of the 3 time horizons through Sharpe-ratio optimization. Among the sampled candidates\, the 5-year Sharpe solution provides the highest expected return and the strongest reward-to-risk trade-off\, at the cost of the highest volatility across the three horizons \n \n \n \n \n Figure 1. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 2. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 3. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Comparison of the 3 time horizons through CVaR (5%). At N=100\, Mean-CVaR shifts the selected portfolios toward lower-volatility regions. The 5-year solution achieves the highest expected return while slightly reducing volatility relative to the 3-year case\, whereas the 10-year solution is the most conservative in volatility terms. \n \n \n \n \n Figure 4. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 5. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 6. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 7. Sharpe vs CVaR: Cross-Horizon Comparison. Sharpe is higher at all three horizons. CVaR is lower at all three horizons. \n \n \n \n \n \n \n \n \n \n \n Contact: Silvi-Maria Gurova | IICT-BAS | silvimaria.gurova@iict.bas.bgTodor Gurov | IICT-BAS | todor.gurov@iict.bas.bgAleksandar Kirilov | IICT-BAS | alex.kirilov@iict.bas.bg URL:https://eurocc-bulgaria.bg/?tribe_events=molecular-dynamics-simulation-with-gromacs-organized-by-ncc-austria-25-26-mar-2026-online ATTACH;FMTTYPE=image/png:https://eurocc-bulgaria.bg/wp-content/uploads/2026/03/MD-2026-03_Gromax.png END:VEVENT BEGIN:VEVENT DTSTART;VALUE=DATE:20260310 DTEND;VALUE=DATE:20260313 DTSTAMP:20261023T101420 CREATED:20260225T080018Z LAST-MODIFIED:20260304T081318Z UID:7007-1773100800-1773359999@eurocc-bulgaria.bg SUMMARY:EuroHPC Summit 2026\, 10-12 March 2026\, Paphos\, Cyprus DESCRIPTION:NCC BULGARIA\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n NCC presenting the success story: NCC Bulgaria is the National Competence Centre for HPC and HPDA\, operating within the EuroCC3 network and EuroHPC ecosystem. Coordinated by the Institute of Information and Communication Technologies at the Bulgarian Academy of Sciences (IICT-BAS) in partnership with the Sofia University “St. Kliment Ohridski” (SU) and the University of National and World Economy (UNWE)\, it supports industry\, academia\, and public organizations through specialized training\, consulting\, technical expertise\, and access to national and European HPC resources. Scientific partners involved The Institute of Information and Communication Technologies (IICT-BAS) conducts fundamental and applied scientific research in the fields of high-performance computing (HPC)\, machine learning\, big data processing\, cybersecurity\, optimization\, financial modeling and mathematical modeling. The developed methods and technologies find application in medicine\, ecology\, industry\, transport\, financial analysis\, process optimization\, and digital transformation. Technical/scientific Challenge Constructing reliable investment portfolios requires balancing expected returns\, market volatility\, and potential losses under adverse market conditions. For a diversified universe of 32 global financial assets\, the large number of possible allocations makes constrained portfolio optimization computationally demanding. An additional challenge is that different risk measures\, such as the Sharpe ratio and Conditional Value-at-Risk (CVaR)\, may favour substantially different investment strategies. Portfolio decisions are also sensitive to the historical estimation period and the numerical sampling method. The research therefore investigates how Monte Carlo (MC) and randomized Quasi-Monte Carlo (QMC) algorithms can efficiently identify feasible portfolios while assessing numerical accuracy\, computational cost\, and sensitivity to market conditions. \n \n \n \n \n \n \n \n \n \n \n Solution Researchers at the Institute of Information and Communication Technologies of the Bulgarian Academy of Sciences (IICT-BAS) developed a reproducible Python-based framework for stochastic portfolio optimization using the HEMUS supercomputer. The methodology compares classical Monte Carlo sampling with Mersenne Twister and randomized Quasi-Monte Carlo methods based on scrambled Sobol and randomized Halton sequences. Candidate portfolios are generated under full-investment and long-only constraints\, with a maximum allocation of 30% per asset. We investigate two optimization objectives: maximizing the Sharpe ratio to improve risk-adjusted returns and minimizing historical CVaR to limit losses in the worst 5% of monthly observations. The analysis considers three historical estimation windows of 3\, 5\, and 10 years and five candidate budgets ranging from 128 to 2048. We evaluate each configuration through 50 independent randomized repetitions and compare it with deterministic numerical benchmarks using approximation errors\, portfolio characteristics\, and execution times. This computational framework enables systematic analysis of how sampling techniques\, risk objectives and historical periods influence portfolio selection. Scientific impact Methodology and Computational Framework The research presents a reproducible Python-based framework for stochastic portfolio optimization\, implemented on the HEMUS supercomputer at IICT-BAS. The study examines 32 global financial assets across 3-\, 5- and 10-year historical windows\, comparing Monte Carlo (with pseudorandom generator Mersenne Twister) and randomized Quasi-Monte Carlo (with scrambled Sobol and Halton) methods under Sharpe ratio maximization and CVaR minimization. Experiment 1 – Accuracy and Convergence of Stochastic Portfolio Search: Five sampling budgets (128–2048 candidate portfolios) and 50 independent repetitions per configuration were evaluated against deterministic numerical benchmarks. Increasing the number of candidates reduced approximation errors\, while differences between sampling methods remained modest\, with classical Monte Carlo generally requiring less execution time.Experiment 2 – Financial Risk – Return Comparison: The 5-year Sharpe-oriented portfolios achieved a mean annualized return of 9.15%\, while the 10-year CVaR-oriented portfolios recorded the lowest mean historical monthly CVaR of 4.40% and annualized volatility of 6.59%. These results highlight the trade-off between maximizing risk-adjusted returns and minimizing downside risk.Experiment 3 – Historical Window and Sampling Sensitivity: The 3-\, 5- and 10-year estimation windows produced different portfolio allocations and risk–return characteristics. The optimization objective and historical period generally had a greater influence on portfolio selection than the choice of sampling generator.Scientific Contribution The study demonstrates the potential of HPC-supported stochastic optimization for reproducible financial risk analysis and numerical benchmarking. The HEMUS supercomputer enabled extensive computational experiments\, providing a foundation for scalable portfolio optimization and future financial decision-support systems. Benefits:  Improved optimization accuracy: Larger sampling budgets reduce approximation errors relative to deterministic benchmarks.Risk-aware portfolio selection: Compares return-oriented Sharpe optimization with downside-risk-focused CVaR.Reliable numerical benchmarking: Evaluates accuracy\, stability and computational efficiency across MC/RQMC methods.Better understanding of market sensitivity: Reveals how historical estimation windows influence investment outcomes.HPC-enabled research: Supports repeated simulations and scalable numerical experimentation.Future financial applications: Provides a foundation for computational portfolio analysis and investment decision-support systems.\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Success story # Highlights: Keywords: HPC; Monte Carlo; Quasi-Monte Carlo; Portfolio Optimization; Sharpe Ratio; CVaR; Financial Risk.Technology: HEMUS Supercomputer; Python; Mersenne Twister; Scrambled Sobol and Halton; Stochastic Optimization.Industry sectors: Finance/Insurance; IT/HPC Systems\, Services & Software Providers.Scientific achievement: Development of a reproducible HPC-based framework for comparing stochastic portfolio optimization methods.Key results: Larger sampling budgets improved numerical accuracy\, while optimization objectives and historical windows significantly influenced portfolio risk–return characteristics.HPC contribution: HEMUS enabled repeated simulations\, algorithm benchmarking and computational performance analysis.\n \n \n \n \n Comparison of the 3 time horizons through Sharpe-ratio optimization. Among the sampled candidates\, the 5-year Sharpe solution provides the highest expected return and the strongest reward-to-risk trade-off\, at the cost of the highest volatility across the three horizons \n \n \n \n \n Figure 1. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 2. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 3. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Comparison of the 3 time horizons through CVaR (5%). At N=100\, Mean-CVaR shifts the selected portfolios toward lower-volatility regions. The 5-year solution achieves the highest expected return while slightly reducing volatility relative to the 3-year case\, whereas the 10-year solution is the most conservative in volatility terms. \n \n \n \n \n Figure 4. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 5. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 6. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 7. Sharpe vs CVaR: Cross-Horizon Comparison. Sharpe is higher at all three horizons. CVaR is lower at all three horizons. \n \n \n \n \n \n \n \n \n \n \n Contact: Silvi-Maria Gurova | IICT-BAS | silvimaria.gurova@iict.bas.bgTodor Gurov | IICT-BAS | todor.gurov@iict.bas.bgAleksandar Kirilov | IICT-BAS | alex.kirilov@iict.bas.bg URL:https://eurocc-bulgaria.bg/?tribe_events=eurohpc-summit-2026-10-12-march-2026-paphos-cyprus END:VEVENT BEGIN:VEVENT DTSTART;TZID=Europe/Helsinki:20260303T120000 DTEND;TZID=Europe/Helsinki:20260303T133000 DTSTAMP:20261023T101420 CREATED:20260222T090007Z LAST-MODIFIED:20260316T135017Z UID:7065-1772539200-1772544600@eurocc-bulgaria.bg SUMMARY:ENCCS Webinar Learning Electronic Structure from X-ray Spectroscopy: Multiplet Theory\, Data\, and High-Performance Computing\, March 3rd\, 2026\, NCC Sweden\, Online DESCRIPTION:NCC BULGARIA\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n NCC presenting the success story: NCC Bulgaria is the National Competence Centre for HPC and HPDA\, operating within the EuroCC3 network and EuroHPC ecosystem. Coordinated by the Institute of Information and Communication Technologies at the Bulgarian Academy of Sciences (IICT-BAS) in partnership with the Sofia University “St. Kliment Ohridski” (SU) and the University of National and World Economy (UNWE)\, it supports industry\, academia\, and public organizations through specialized training\, consulting\, technical expertise\, and access to national and European HPC resources. Scientific partners involved The Institute of Information and Communication Technologies (IICT-BAS) conducts fundamental and applied scientific research in the fields of high-performance computing (HPC)\, machine learning\, big data processing\, cybersecurity\, optimization\, financial modeling and mathematical modeling. The developed methods and technologies find application in medicine\, ecology\, industry\, transport\, financial analysis\, process optimization\, and digital transformation. Technical/scientific Challenge Constructing reliable investment portfolios requires balancing expected returns\, market volatility\, and potential losses under adverse market conditions. For a diversified universe of 32 global financial assets\, the large number of possible allocations makes constrained portfolio optimization computationally demanding. An additional challenge is that different risk measures\, such as the Sharpe ratio and Conditional Value-at-Risk (CVaR)\, may favour substantially different investment strategies. Portfolio decisions are also sensitive to the historical estimation period and the numerical sampling method. The research therefore investigates how Monte Carlo (MC) and randomized Quasi-Monte Carlo (QMC) algorithms can efficiently identify feasible portfolios while assessing numerical accuracy\, computational cost\, and sensitivity to market conditions. \n \n \n \n \n \n \n \n \n \n \n Solution Researchers at the Institute of Information and Communication Technologies of the Bulgarian Academy of Sciences (IICT-BAS) developed a reproducible Python-based framework for stochastic portfolio optimization using the HEMUS supercomputer. The methodology compares classical Monte Carlo sampling with Mersenne Twister and randomized Quasi-Monte Carlo methods based on scrambled Sobol and randomized Halton sequences. Candidate portfolios are generated under full-investment and long-only constraints\, with a maximum allocation of 30% per asset. We investigate two optimization objectives: maximizing the Sharpe ratio to improve risk-adjusted returns and minimizing historical CVaR to limit losses in the worst 5% of monthly observations. The analysis considers three historical estimation windows of 3\, 5\, and 10 years and five candidate budgets ranging from 128 to 2048. We evaluate each configuration through 50 independent randomized repetitions and compare it with deterministic numerical benchmarks using approximation errors\, portfolio characteristics\, and execution times. This computational framework enables systematic analysis of how sampling techniques\, risk objectives and historical periods influence portfolio selection. Scientific impact Methodology and Computational Framework The research presents a reproducible Python-based framework for stochastic portfolio optimization\, implemented on the HEMUS supercomputer at IICT-BAS. The study examines 32 global financial assets across 3-\, 5- and 10-year historical windows\, comparing Monte Carlo (with pseudorandom generator Mersenne Twister) and randomized Quasi-Monte Carlo (with scrambled Sobol and Halton) methods under Sharpe ratio maximization and CVaR minimization. Experiment 1 – Accuracy and Convergence of Stochastic Portfolio Search: Five sampling budgets (128–2048 candidate portfolios) and 50 independent repetitions per configuration were evaluated against deterministic numerical benchmarks. Increasing the number of candidates reduced approximation errors\, while differences between sampling methods remained modest\, with classical Monte Carlo generally requiring less execution time.Experiment 2 – Financial Risk – Return Comparison: The 5-year Sharpe-oriented portfolios achieved a mean annualized return of 9.15%\, while the 10-year CVaR-oriented portfolios recorded the lowest mean historical monthly CVaR of 4.40% and annualized volatility of 6.59%. These results highlight the trade-off between maximizing risk-adjusted returns and minimizing downside risk.Experiment 3 – Historical Window and Sampling Sensitivity: The 3-\, 5- and 10-year estimation windows produced different portfolio allocations and risk–return characteristics. The optimization objective and historical period generally had a greater influence on portfolio selection than the choice of sampling generator.Scientific Contribution The study demonstrates the potential of HPC-supported stochastic optimization for reproducible financial risk analysis and numerical benchmarking. The HEMUS supercomputer enabled extensive computational experiments\, providing a foundation for scalable portfolio optimization and future financial decision-support systems. Benefits:  Improved optimization accuracy: Larger sampling budgets reduce approximation errors relative to deterministic benchmarks.Risk-aware portfolio selection: Compares return-oriented Sharpe optimization with downside-risk-focused CVaR.Reliable numerical benchmarking: Evaluates accuracy\, stability and computational efficiency across MC/RQMC methods.Better understanding of market sensitivity: Reveals how historical estimation windows influence investment outcomes.HPC-enabled research: Supports repeated simulations and scalable numerical experimentation.Future financial applications: Provides a foundation for computational portfolio analysis and investment decision-support systems.\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Success story # Highlights: Keywords: HPC; Monte Carlo; Quasi-Monte Carlo; Portfolio Optimization; Sharpe Ratio; CVaR; Financial Risk.Technology: HEMUS Supercomputer; Python; Mersenne Twister; Scrambled Sobol and Halton; Stochastic Optimization.Industry sectors: Finance/Insurance; IT/HPC Systems\, Services & Software Providers.Scientific achievement: Development of a reproducible HPC-based framework for comparing stochastic portfolio optimization methods.Key results: Larger sampling budgets improved numerical accuracy\, while optimization objectives and historical windows significantly influenced portfolio risk–return characteristics.HPC contribution: HEMUS enabled repeated simulations\, algorithm benchmarking and computational performance analysis.\n \n \n \n \n Comparison of the 3 time horizons through Sharpe-ratio optimization. Among the sampled candidates\, the 5-year Sharpe solution provides the highest expected return and the strongest reward-to-risk trade-off\, at the cost of the highest volatility across the three horizons \n \n \n \n \n Figure 1. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 2. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 3. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Comparison of the 3 time horizons through CVaR (5%). At N=100\, Mean-CVaR shifts the selected portfolios toward lower-volatility regions. The 5-year solution achieves the highest expected return while slightly reducing volatility relative to the 3-year case\, whereas the 10-year solution is the most conservative in volatility terms. \n \n \n \n \n Figure 4. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 5. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 6. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 7. Sharpe vs CVaR: Cross-Horizon Comparison. Sharpe is higher at all three horizons. CVaR is lower at all three horizons. \n \n \n \n \n \n \n \n \n \n \n Contact: Silvi-Maria Gurova | IICT-BAS | silvimaria.gurova@iict.bas.bgTodor Gurov | IICT-BAS | todor.gurov@iict.bas.bgAleksandar Kirilov | IICT-BAS | alex.kirilov@iict.bas.bg URL:https://eurocc-bulgaria.bg/?tribe_events=enccs-webinar-learning-electronic-structure-from-x-ray-spectroscopy-multiplet-theory-data-and-high-performance-computing-march-3rd-2026-ncc-sweden-online ATTACH;FMTTYPE=image/jpeg:https://eurocc-bulgaria.bg/wp-content/uploads/2026/03/EnccsWebinar26.jpg END:VEVENT BEGIN:VEVENT DTSTART;TZID=Europe/Helsinki:20260302T093000 DTEND;TZID=Europe/Helsinki:20260303T123000 DTSTAMP:20261023T101420 CREATED:20260216T080020Z LAST-MODIFIED:20260319T083742Z UID:7120-1772443800-1772541000@eurocc-bulgaria.bg SUMMARY:AI in Action for SMEs\, 2–3 March 2026\, organized by NCC Montenegro\, NCC Serbia\, NCC Bosnia and Herzegovina\, and NCC Türkiye\, online DESCRIPTION:NCC BULGARIA\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n NCC presenting the success story: NCC Bulgaria is the National Competence Centre for HPC and HPDA\, operating within the EuroCC3 network and EuroHPC ecosystem. Coordinated by the Institute of Information and Communication Technologies at the Bulgarian Academy of Sciences (IICT-BAS) in partnership with the Sofia University “St. Kliment Ohridski” (SU) and the University of National and World Economy (UNWE)\, it supports industry\, academia\, and public organizations through specialized training\, consulting\, technical expertise\, and access to national and European HPC resources. Scientific partners involved The Institute of Information and Communication Technologies (IICT-BAS) conducts fundamental and applied scientific research in the fields of high-performance computing (HPC)\, machine learning\, big data processing\, cybersecurity\, optimization\, financial modeling and mathematical modeling. The developed methods and technologies find application in medicine\, ecology\, industry\, transport\, financial analysis\, process optimization\, and digital transformation. Technical/scientific Challenge Constructing reliable investment portfolios requires balancing expected returns\, market volatility\, and potential losses under adverse market conditions. For a diversified universe of 32 global financial assets\, the large number of possible allocations makes constrained portfolio optimization computationally demanding. An additional challenge is that different risk measures\, such as the Sharpe ratio and Conditional Value-at-Risk (CVaR)\, may favour substantially different investment strategies. Portfolio decisions are also sensitive to the historical estimation period and the numerical sampling method. The research therefore investigates how Monte Carlo (MC) and randomized Quasi-Monte Carlo (QMC) algorithms can efficiently identify feasible portfolios while assessing numerical accuracy\, computational cost\, and sensitivity to market conditions. \n \n \n \n \n \n \n \n \n \n \n Solution Researchers at the Institute of Information and Communication Technologies of the Bulgarian Academy of Sciences (IICT-BAS) developed a reproducible Python-based framework for stochastic portfolio optimization using the HEMUS supercomputer. The methodology compares classical Monte Carlo sampling with Mersenne Twister and randomized Quasi-Monte Carlo methods based on scrambled Sobol and randomized Halton sequences. Candidate portfolios are generated under full-investment and long-only constraints\, with a maximum allocation of 30% per asset. We investigate two optimization objectives: maximizing the Sharpe ratio to improve risk-adjusted returns and minimizing historical CVaR to limit losses in the worst 5% of monthly observations. The analysis considers three historical estimation windows of 3\, 5\, and 10 years and five candidate budgets ranging from 128 to 2048. We evaluate each configuration through 50 independent randomized repetitions and compare it with deterministic numerical benchmarks using approximation errors\, portfolio characteristics\, and execution times. This computational framework enables systematic analysis of how sampling techniques\, risk objectives and historical periods influence portfolio selection. Scientific impact Methodology and Computational Framework The research presents a reproducible Python-based framework for stochastic portfolio optimization\, implemented on the HEMUS supercomputer at IICT-BAS. The study examines 32 global financial assets across 3-\, 5- and 10-year historical windows\, comparing Monte Carlo (with pseudorandom generator Mersenne Twister) and randomized Quasi-Monte Carlo (with scrambled Sobol and Halton) methods under Sharpe ratio maximization and CVaR minimization. Experiment 1 – Accuracy and Convergence of Stochastic Portfolio Search: Five sampling budgets (128–2048 candidate portfolios) and 50 independent repetitions per configuration were evaluated against deterministic numerical benchmarks. Increasing the number of candidates reduced approximation errors\, while differences between sampling methods remained modest\, with classical Monte Carlo generally requiring less execution time.Experiment 2 – Financial Risk – Return Comparison: The 5-year Sharpe-oriented portfolios achieved a mean annualized return of 9.15%\, while the 10-year CVaR-oriented portfolios recorded the lowest mean historical monthly CVaR of 4.40% and annualized volatility of 6.59%. These results highlight the trade-off between maximizing risk-adjusted returns and minimizing downside risk.Experiment 3 – Historical Window and Sampling Sensitivity: The 3-\, 5- and 10-year estimation windows produced different portfolio allocations and risk–return characteristics. The optimization objective and historical period generally had a greater influence on portfolio selection than the choice of sampling generator.Scientific Contribution The study demonstrates the potential of HPC-supported stochastic optimization for reproducible financial risk analysis and numerical benchmarking. The HEMUS supercomputer enabled extensive computational experiments\, providing a foundation for scalable portfolio optimization and future financial decision-support systems. Benefits:  Improved optimization accuracy: Larger sampling budgets reduce approximation errors relative to deterministic benchmarks.Risk-aware portfolio selection: Compares return-oriented Sharpe optimization with downside-risk-focused CVaR.Reliable numerical benchmarking: Evaluates accuracy\, stability and computational efficiency across MC/RQMC methods.Better understanding of market sensitivity: Reveals how historical estimation windows influence investment outcomes.HPC-enabled research: Supports repeated simulations and scalable numerical experimentation.Future financial applications: Provides a foundation for computational portfolio analysis and investment decision-support systems.\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Success story # Highlights: Keywords: HPC; Monte Carlo; Quasi-Monte Carlo; Portfolio Optimization; Sharpe Ratio; CVaR; Financial Risk.Technology: HEMUS Supercomputer; Python; Mersenne Twister; Scrambled Sobol and Halton; Stochastic Optimization.Industry sectors: Finance/Insurance; IT/HPC Systems\, Services & Software Providers.Scientific achievement: Development of a reproducible HPC-based framework for comparing stochastic portfolio optimization methods.Key results: Larger sampling budgets improved numerical accuracy\, while optimization objectives and historical windows significantly influenced portfolio risk–return characteristics.HPC contribution: HEMUS enabled repeated simulations\, algorithm benchmarking and computational performance analysis.\n \n \n \n \n Comparison of the 3 time horizons through Sharpe-ratio optimization. Among the sampled candidates\, the 5-year Sharpe solution provides the highest expected return and the strongest reward-to-risk trade-off\, at the cost of the highest volatility across the three horizons \n \n \n \n \n Figure 1. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 2. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 3. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Comparison of the 3 time horizons through CVaR (5%). At N=100\, Mean-CVaR shifts the selected portfolios toward lower-volatility regions. The 5-year solution achieves the highest expected return while slightly reducing volatility relative to the 3-year case\, whereas the 10-year solution is the most conservative in volatility terms. \n \n \n \n \n Figure 4. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 5. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 6. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 7. Sharpe vs CVaR: Cross-Horizon Comparison. Sharpe is higher at all three horizons. CVaR is lower at all three horizons. \n \n \n \n \n \n \n \n \n \n \n Contact: Silvi-Maria Gurova | IICT-BAS | silvimaria.gurova@iict.bas.bgTodor Gurov | IICT-BAS | todor.gurov@iict.bas.bgAleksandar Kirilov | IICT-BAS | alex.kirilov@iict.bas.bg URL:https://eurocc-bulgaria.bg/?tribe_events=ai-in-action-for-smes-2-3-march-2026-organized-by-ncc-montenegro-ncc-serbia-ncc-bosnia-and-herzegovina-and-ncc-turkiye-online ATTACH;FMTTYPE=image/png:https://eurocc-bulgaria.bg/wp-content/uploads/2026/03/EuroCC4SEEcolabPoster26.png END:VEVENT BEGIN:VEVENT DTSTART;VALUE=DATE:20260226 DTEND;VALUE=DATE:20260227 DTSTAMP:20261023T101420 CREATED:20260213T090015Z LAST-MODIFIED:20260416T124713Z UID:7155-1772064000-1772150399@eurocc-bulgaria.bg SUMMARY:Consultation for Bulgarian SME on FFplus Third Calls for Business Experiments and Innovation Studies\, 26 February\, 2026 DESCRIPTION:NCC BULGARIA\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n NCC presenting the success story: NCC Bulgaria is the National Competence Centre for HPC and HPDA\, operating within the EuroCC3 network and EuroHPC ecosystem. Coordinated by the Institute of Information and Communication Technologies at the Bulgarian Academy of Sciences (IICT-BAS) in partnership with the Sofia University “St. Kliment Ohridski” (SU) and the University of National and World Economy (UNWE)\, it supports industry\, academia\, and public organizations through specialized training\, consulting\, technical expertise\, and access to national and European HPC resources. Scientific partners involved The Institute of Information and Communication Technologies (IICT-BAS) conducts fundamental and applied scientific research in the fields of high-performance computing (HPC)\, machine learning\, big data processing\, cybersecurity\, optimization\, financial modeling and mathematical modeling. The developed methods and technologies find application in medicine\, ecology\, industry\, transport\, financial analysis\, process optimization\, and digital transformation. Technical/scientific Challenge Constructing reliable investment portfolios requires balancing expected returns\, market volatility\, and potential losses under adverse market conditions. For a diversified universe of 32 global financial assets\, the large number of possible allocations makes constrained portfolio optimization computationally demanding. An additional challenge is that different risk measures\, such as the Sharpe ratio and Conditional Value-at-Risk (CVaR)\, may favour substantially different investment strategies. Portfolio decisions are also sensitive to the historical estimation period and the numerical sampling method. The research therefore investigates how Monte Carlo (MC) and randomized Quasi-Monte Carlo (QMC) algorithms can efficiently identify feasible portfolios while assessing numerical accuracy\, computational cost\, and sensitivity to market conditions. \n \n \n \n \n \n \n \n \n \n \n Solution Researchers at the Institute of Information and Communication Technologies of the Bulgarian Academy of Sciences (IICT-BAS) developed a reproducible Python-based framework for stochastic portfolio optimization using the HEMUS supercomputer. The methodology compares classical Monte Carlo sampling with Mersenne Twister and randomized Quasi-Monte Carlo methods based on scrambled Sobol and randomized Halton sequences. Candidate portfolios are generated under full-investment and long-only constraints\, with a maximum allocation of 30% per asset. We investigate two optimization objectives: maximizing the Sharpe ratio to improve risk-adjusted returns and minimizing historical CVaR to limit losses in the worst 5% of monthly observations. The analysis considers three historical estimation windows of 3\, 5\, and 10 years and five candidate budgets ranging from 128 to 2048. We evaluate each configuration through 50 independent randomized repetitions and compare it with deterministic numerical benchmarks using approximation errors\, portfolio characteristics\, and execution times. This computational framework enables systematic analysis of how sampling techniques\, risk objectives and historical periods influence portfolio selection. Scientific impact Methodology and Computational Framework The research presents a reproducible Python-based framework for stochastic portfolio optimization\, implemented on the HEMUS supercomputer at IICT-BAS. The study examines 32 global financial assets across 3-\, 5- and 10-year historical windows\, comparing Monte Carlo (with pseudorandom generator Mersenne Twister) and randomized Quasi-Monte Carlo (with scrambled Sobol and Halton) methods under Sharpe ratio maximization and CVaR minimization. Experiment 1 – Accuracy and Convergence of Stochastic Portfolio Search: Five sampling budgets (128–2048 candidate portfolios) and 50 independent repetitions per configuration were evaluated against deterministic numerical benchmarks. Increasing the number of candidates reduced approximation errors\, while differences between sampling methods remained modest\, with classical Monte Carlo generally requiring less execution time.Experiment 2 – Financial Risk – Return Comparison: The 5-year Sharpe-oriented portfolios achieved a mean annualized return of 9.15%\, while the 10-year CVaR-oriented portfolios recorded the lowest mean historical monthly CVaR of 4.40% and annualized volatility of 6.59%. These results highlight the trade-off between maximizing risk-adjusted returns and minimizing downside risk.Experiment 3 – Historical Window and Sampling Sensitivity: The 3-\, 5- and 10-year estimation windows produced different portfolio allocations and risk–return characteristics. The optimization objective and historical period generally had a greater influence on portfolio selection than the choice of sampling generator.Scientific Contribution The study demonstrates the potential of HPC-supported stochastic optimization for reproducible financial risk analysis and numerical benchmarking. The HEMUS supercomputer enabled extensive computational experiments\, providing a foundation for scalable portfolio optimization and future financial decision-support systems. Benefits:  Improved optimization accuracy: Larger sampling budgets reduce approximation errors relative to deterministic benchmarks.Risk-aware portfolio selection: Compares return-oriented Sharpe optimization with downside-risk-focused CVaR.Reliable numerical benchmarking: Evaluates accuracy\, stability and computational efficiency across MC/RQMC methods.Better understanding of market sensitivity: Reveals how historical estimation windows influence investment outcomes.HPC-enabled research: Supports repeated simulations and scalable numerical experimentation.Future financial applications: Provides a foundation for computational portfolio analysis and investment decision-support systems.\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Success story # Highlights: Keywords: HPC; Monte Carlo; Quasi-Monte Carlo; Portfolio Optimization; Sharpe Ratio; CVaR; Financial Risk.Technology: HEMUS Supercomputer; Python; Mersenne Twister; Scrambled Sobol and Halton; Stochastic Optimization.Industry sectors: Finance/Insurance; IT/HPC Systems\, Services & Software Providers.Scientific achievement: Development of a reproducible HPC-based framework for comparing stochastic portfolio optimization methods.Key results: Larger sampling budgets improved numerical accuracy\, while optimization objectives and historical windows significantly influenced portfolio risk–return characteristics.HPC contribution: HEMUS enabled repeated simulations\, algorithm benchmarking and computational performance analysis.\n \n \n \n \n Comparison of the 3 time horizons through Sharpe-ratio optimization. Among the sampled candidates\, the 5-year Sharpe solution provides the highest expected return and the strongest reward-to-risk trade-off\, at the cost of the highest volatility across the three horizons \n \n \n \n \n Figure 1. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 2. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 3. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Comparison of the 3 time horizons through CVaR (5%). At N=100\, Mean-CVaR shifts the selected portfolios toward lower-volatility regions. The 5-year solution achieves the highest expected return while slightly reducing volatility relative to the 3-year case\, whereas the 10-year solution is the most conservative in volatility terms. \n \n \n \n \n Figure 4. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 5. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 6. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 7. Sharpe vs CVaR: Cross-Horizon Comparison. Sharpe is higher at all three horizons. CVaR is lower at all three horizons. \n \n \n \n \n \n \n \n \n \n \n Contact: Silvi-Maria Gurova | IICT-BAS | silvimaria.gurova@iict.bas.bgTodor Gurov | IICT-BAS | todor.gurov@iict.bas.bgAleksandar Kirilov | IICT-BAS | alex.kirilov@iict.bas.bg URL:https://eurocc-bulgaria.bg/?tribe_events=consultation-for-bulgarian-sme-on-ffplus-third-calls-for-business-experiments-and-innovation-studies-26-february-2026 END:VEVENT BEGIN:VEVENT DTSTART;TZID=Europe/Helsinki:20260224T080000 DTEND;TZID=Europe/Helsinki:20260224T170000 DTSTAMP:20261023T101420 CREATED:20260202T110002Z LAST-MODIFIED:20260402T113438Z UID:7141-1771920000-1771952400@eurocc-bulgaria.bg SUMMARY:HPC seminar series #7 "Parallel stochastic computations for matrix-vector products" DESCRIPTION:NCC BULGARIA\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n NCC presenting the success story: NCC Bulgaria is the National Competence Centre for HPC and HPDA\, operating within the EuroCC3 network and EuroHPC ecosystem. Coordinated by the Institute of Information and Communication Technologies at the Bulgarian Academy of Sciences (IICT-BAS) in partnership with the Sofia University “St. Kliment Ohridski” (SU) and the University of National and World Economy (UNWE)\, it supports industry\, academia\, and public organizations through specialized training\, consulting\, technical expertise\, and access to national and European HPC resources. Scientific partners involved The Institute of Information and Communication Technologies (IICT-BAS) conducts fundamental and applied scientific research in the fields of high-performance computing (HPC)\, machine learning\, big data processing\, cybersecurity\, optimization\, financial modeling and mathematical modeling. The developed methods and technologies find application in medicine\, ecology\, industry\, transport\, financial analysis\, process optimization\, and digital transformation. Technical/scientific Challenge Constructing reliable investment portfolios requires balancing expected returns\, market volatility\, and potential losses under adverse market conditions. For a diversified universe of 32 global financial assets\, the large number of possible allocations makes constrained portfolio optimization computationally demanding. An additional challenge is that different risk measures\, such as the Sharpe ratio and Conditional Value-at-Risk (CVaR)\, may favour substantially different investment strategies. Portfolio decisions are also sensitive to the historical estimation period and the numerical sampling method. The research therefore investigates how Monte Carlo (MC) and randomized Quasi-Monte Carlo (QMC) algorithms can efficiently identify feasible portfolios while assessing numerical accuracy\, computational cost\, and sensitivity to market conditions. \n \n \n \n \n \n \n \n \n \n \n Solution Researchers at the Institute of Information and Communication Technologies of the Bulgarian Academy of Sciences (IICT-BAS) developed a reproducible Python-based framework for stochastic portfolio optimization using the HEMUS supercomputer. The methodology compares classical Monte Carlo sampling with Mersenne Twister and randomized Quasi-Monte Carlo methods based on scrambled Sobol and randomized Halton sequences. Candidate portfolios are generated under full-investment and long-only constraints\, with a maximum allocation of 30% per asset. We investigate two optimization objectives: maximizing the Sharpe ratio to improve risk-adjusted returns and minimizing historical CVaR to limit losses in the worst 5% of monthly observations. The analysis considers three historical estimation windows of 3\, 5\, and 10 years and five candidate budgets ranging from 128 to 2048. We evaluate each configuration through 50 independent randomized repetitions and compare it with deterministic numerical benchmarks using approximation errors\, portfolio characteristics\, and execution times. This computational framework enables systematic analysis of how sampling techniques\, risk objectives and historical periods influence portfolio selection. Scientific impact Methodology and Computational Framework The research presents a reproducible Python-based framework for stochastic portfolio optimization\, implemented on the HEMUS supercomputer at IICT-BAS. The study examines 32 global financial assets across 3-\, 5- and 10-year historical windows\, comparing Monte Carlo (with pseudorandom generator Mersenne Twister) and randomized Quasi-Monte Carlo (with scrambled Sobol and Halton) methods under Sharpe ratio maximization and CVaR minimization. Experiment 1 – Accuracy and Convergence of Stochastic Portfolio Search: Five sampling budgets (128–2048 candidate portfolios) and 50 independent repetitions per configuration were evaluated against deterministic numerical benchmarks. Increasing the number of candidates reduced approximation errors\, while differences between sampling methods remained modest\, with classical Monte Carlo generally requiring less execution time.Experiment 2 – Financial Risk – Return Comparison: The 5-year Sharpe-oriented portfolios achieved a mean annualized return of 9.15%\, while the 10-year CVaR-oriented portfolios recorded the lowest mean historical monthly CVaR of 4.40% and annualized volatility of 6.59%. These results highlight the trade-off between maximizing risk-adjusted returns and minimizing downside risk.Experiment 3 – Historical Window and Sampling Sensitivity: The 3-\, 5- and 10-year estimation windows produced different portfolio allocations and risk–return characteristics. The optimization objective and historical period generally had a greater influence on portfolio selection than the choice of sampling generator.Scientific Contribution The study demonstrates the potential of HPC-supported stochastic optimization for reproducible financial risk analysis and numerical benchmarking. The HEMUS supercomputer enabled extensive computational experiments\, providing a foundation for scalable portfolio optimization and future financial decision-support systems. Benefits:  Improved optimization accuracy: Larger sampling budgets reduce approximation errors relative to deterministic benchmarks.Risk-aware portfolio selection: Compares return-oriented Sharpe optimization with downside-risk-focused CVaR.Reliable numerical benchmarking: Evaluates accuracy\, stability and computational efficiency across MC/RQMC methods.Better understanding of market sensitivity: Reveals how historical estimation windows influence investment outcomes.HPC-enabled research: Supports repeated simulations and scalable numerical experimentation.Future financial applications: Provides a foundation for computational portfolio analysis and investment decision-support systems.\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Success story # Highlights: Keywords: HPC; Monte Carlo; Quasi-Monte Carlo; Portfolio Optimization; Sharpe Ratio; CVaR; Financial Risk.Technology: HEMUS Supercomputer; Python; Mersenne Twister; Scrambled Sobol and Halton; Stochastic Optimization.Industry sectors: Finance/Insurance; IT/HPC Systems\, Services & Software Providers.Scientific achievement: Development of a reproducible HPC-based framework for comparing stochastic portfolio optimization methods.Key results: Larger sampling budgets improved numerical accuracy\, while optimization objectives and historical windows significantly influenced portfolio risk–return characteristics.HPC contribution: HEMUS enabled repeated simulations\, algorithm benchmarking and computational performance analysis.\n \n \n \n \n Comparison of the 3 time horizons through Sharpe-ratio optimization. Among the sampled candidates\, the 5-year Sharpe solution provides the highest expected return and the strongest reward-to-risk trade-off\, at the cost of the highest volatility across the three horizons \n \n \n \n \n Figure 1. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 2. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 3. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Comparison of the 3 time horizons through CVaR (5%). At N=100\, Mean-CVaR shifts the selected portfolios toward lower-volatility regions. The 5-year solution achieves the highest expected return while slightly reducing volatility relative to the 3-year case\, whereas the 10-year solution is the most conservative in volatility terms. \n \n \n \n \n Figure 4. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 5. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 6. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 7. Sharpe vs CVaR: Cross-Horizon Comparison. Sharpe is higher at all three horizons. CVaR is lower at all three horizons. \n \n \n \n \n \n \n \n \n \n \n Contact: Silvi-Maria Gurova | IICT-BAS | silvimaria.gurova@iict.bas.bgTodor Gurov | IICT-BAS | todor.gurov@iict.bas.bgAleksandar Kirilov | IICT-BAS | alex.kirilov@iict.bas.bg URL:https://eurocc-bulgaria.bg/?tribe_events=hpc-seminar-series-7-parallel-stochastic-computations-for-matrix-vector-products END:VEVENT BEGIN:VEVENT DTSTART;VALUE=DATE:20260209 DTEND;VALUE=DATE:20260214 DTSTAMP:20261023T101420 CREATED:20260202T080038Z LAST-MODIFIED:20260319T083355Z UID:7114-1770595200-1771027199@eurocc-bulgaria.bg SUMMARY:VI-HPS Tuning Workshop 9-13 February\, hybrid\, BSC and online DESCRIPTION:NCC BULGARIA\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n NCC presenting the success story: NCC Bulgaria is the National Competence Centre for HPC and HPDA\, operating within the EuroCC3 network and EuroHPC ecosystem. Coordinated by the Institute of Information and Communication Technologies at the Bulgarian Academy of Sciences (IICT-BAS) in partnership with the Sofia University “St. Kliment Ohridski” (SU) and the University of National and World Economy (UNWE)\, it supports industry\, academia\, and public organizations through specialized training\, consulting\, technical expertise\, and access to national and European HPC resources. Scientific partners involved The Institute of Information and Communication Technologies (IICT-BAS) conducts fundamental and applied scientific research in the fields of high-performance computing (HPC)\, machine learning\, big data processing\, cybersecurity\, optimization\, financial modeling and mathematical modeling. The developed methods and technologies find application in medicine\, ecology\, industry\, transport\, financial analysis\, process optimization\, and digital transformation. Technical/scientific Challenge Constructing reliable investment portfolios requires balancing expected returns\, market volatility\, and potential losses under adverse market conditions. For a diversified universe of 32 global financial assets\, the large number of possible allocations makes constrained portfolio optimization computationally demanding. An additional challenge is that different risk measures\, such as the Sharpe ratio and Conditional Value-at-Risk (CVaR)\, may favour substantially different investment strategies. Portfolio decisions are also sensitive to the historical estimation period and the numerical sampling method. The research therefore investigates how Monte Carlo (MC) and randomized Quasi-Monte Carlo (QMC) algorithms can efficiently identify feasible portfolios while assessing numerical accuracy\, computational cost\, and sensitivity to market conditions. \n \n \n \n \n \n \n \n \n \n \n Solution Researchers at the Institute of Information and Communication Technologies of the Bulgarian Academy of Sciences (IICT-BAS) developed a reproducible Python-based framework for stochastic portfolio optimization using the HEMUS supercomputer. The methodology compares classical Monte Carlo sampling with Mersenne Twister and randomized Quasi-Monte Carlo methods based on scrambled Sobol and randomized Halton sequences. Candidate portfolios are generated under full-investment and long-only constraints\, with a maximum allocation of 30% per asset. We investigate two optimization objectives: maximizing the Sharpe ratio to improve risk-adjusted returns and minimizing historical CVaR to limit losses in the worst 5% of monthly observations. The analysis considers three historical estimation windows of 3\, 5\, and 10 years and five candidate budgets ranging from 128 to 2048. We evaluate each configuration through 50 independent randomized repetitions and compare it with deterministic numerical benchmarks using approximation errors\, portfolio characteristics\, and execution times. This computational framework enables systematic analysis of how sampling techniques\, risk objectives and historical periods influence portfolio selection. Scientific impact Methodology and Computational Framework The research presents a reproducible Python-based framework for stochastic portfolio optimization\, implemented on the HEMUS supercomputer at IICT-BAS. The study examines 32 global financial assets across 3-\, 5- and 10-year historical windows\, comparing Monte Carlo (with pseudorandom generator Mersenne Twister) and randomized Quasi-Monte Carlo (with scrambled Sobol and Halton) methods under Sharpe ratio maximization and CVaR minimization. Experiment 1 – Accuracy and Convergence of Stochastic Portfolio Search: Five sampling budgets (128–2048 candidate portfolios) and 50 independent repetitions per configuration were evaluated against deterministic numerical benchmarks. Increasing the number of candidates reduced approximation errors\, while differences between sampling methods remained modest\, with classical Monte Carlo generally requiring less execution time.Experiment 2 – Financial Risk – Return Comparison: The 5-year Sharpe-oriented portfolios achieved a mean annualized return of 9.15%\, while the 10-year CVaR-oriented portfolios recorded the lowest mean historical monthly CVaR of 4.40% and annualized volatility of 6.59%. These results highlight the trade-off between maximizing risk-adjusted returns and minimizing downside risk.Experiment 3 – Historical Window and Sampling Sensitivity: The 3-\, 5- and 10-year estimation windows produced different portfolio allocations and risk–return characteristics. The optimization objective and historical period generally had a greater influence on portfolio selection than the choice of sampling generator.Scientific Contribution The study demonstrates the potential of HPC-supported stochastic optimization for reproducible financial risk analysis and numerical benchmarking. The HEMUS supercomputer enabled extensive computational experiments\, providing a foundation for scalable portfolio optimization and future financial decision-support systems. Benefits:  Improved optimization accuracy: Larger sampling budgets reduce approximation errors relative to deterministic benchmarks.Risk-aware portfolio selection: Compares return-oriented Sharpe optimization with downside-risk-focused CVaR.Reliable numerical benchmarking: Evaluates accuracy\, stability and computational efficiency across MC/RQMC methods.Better understanding of market sensitivity: Reveals how historical estimation windows influence investment outcomes.HPC-enabled research: Supports repeated simulations and scalable numerical experimentation.Future financial applications: Provides a foundation for computational portfolio analysis and investment decision-support systems.\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Success story # Highlights: Keywords: HPC; Monte Carlo; Quasi-Monte Carlo; Portfolio Optimization; Sharpe Ratio; CVaR; Financial Risk.Technology: HEMUS Supercomputer; Python; Mersenne Twister; Scrambled Sobol and Halton; Stochastic Optimization.Industry sectors: Finance/Insurance; IT/HPC Systems\, Services & Software Providers.Scientific achievement: Development of a reproducible HPC-based framework for comparing stochastic portfolio optimization methods.Key results: Larger sampling budgets improved numerical accuracy\, while optimization objectives and historical windows significantly influenced portfolio risk–return characteristics.HPC contribution: HEMUS enabled repeated simulations\, algorithm benchmarking and computational performance analysis.\n \n \n \n \n Comparison of the 3 time horizons through Sharpe-ratio optimization. Among the sampled candidates\, the 5-year Sharpe solution provides the highest expected return and the strongest reward-to-risk trade-off\, at the cost of the highest volatility across the three horizons \n \n \n \n \n Figure 1. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 2. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 3. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Comparison of the 3 time horizons through CVaR (5%). At N=100\, Mean-CVaR shifts the selected portfolios toward lower-volatility regions. The 5-year solution achieves the highest expected return while slightly reducing volatility relative to the 3-year case\, whereas the 10-year solution is the most conservative in volatility terms. \n \n \n \n \n Figure 4. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 5. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 6. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 7. Sharpe vs CVaR: Cross-Horizon Comparison. Sharpe is higher at all three horizons. CVaR is lower at all three horizons. \n \n \n \n \n \n \n \n \n \n \n Contact: Silvi-Maria Gurova | IICT-BAS | silvimaria.gurova@iict.bas.bgTodor Gurov | IICT-BAS | todor.gurov@iict.bas.bgAleksandar Kirilov | IICT-BAS | alex.kirilov@iict.bas.bg URL:https://eurocc-bulgaria.bg/?tribe_events=vi-hps-tuning-workshop-9-13-february-hybrid-bsc-and-online ATTACH;FMTTYPE=image/jpeg:https://eurocc-bulgaria.bg/wp-content/uploads/2026/03/vi-hps_tuning_workshop_croped_v2.jpg END:VEVENT BEGIN:VEVENT DTSTART;VALUE=DATE:20260205 DTEND;VALUE=DATE:20260206 DTSTAMP:20261023T101420 CREATED:20260121T090000Z LAST-MODIFIED:20260416T124116Z UID:7151-1770249600-1770335999@eurocc-bulgaria.bg SUMMARY:Doctoral Information Meeting On February 5\, 2026 DESCRIPTION:NCC BULGARIA\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n NCC presenting the success story: NCC Bulgaria is the National Competence Centre for HPC and HPDA\, operating within the EuroCC3 network and EuroHPC ecosystem. Coordinated by the Institute of Information and Communication Technologies at the Bulgarian Academy of Sciences (IICT-BAS) in partnership with the Sofia University “St. Kliment Ohridski” (SU) and the University of National and World Economy (UNWE)\, it supports industry\, academia\, and public organizations through specialized training\, consulting\, technical expertise\, and access to national and European HPC resources. Scientific partners involved The Institute of Information and Communication Technologies (IICT-BAS) conducts fundamental and applied scientific research in the fields of high-performance computing (HPC)\, machine learning\, big data processing\, cybersecurity\, optimization\, financial modeling and mathematical modeling. The developed methods and technologies find application in medicine\, ecology\, industry\, transport\, financial analysis\, process optimization\, and digital transformation. Technical/scientific Challenge Constructing reliable investment portfolios requires balancing expected returns\, market volatility\, and potential losses under adverse market conditions. For a diversified universe of 32 global financial assets\, the large number of possible allocations makes constrained portfolio optimization computationally demanding. An additional challenge is that different risk measures\, such as the Sharpe ratio and Conditional Value-at-Risk (CVaR)\, may favour substantially different investment strategies. Portfolio decisions are also sensitive to the historical estimation period and the numerical sampling method. The research therefore investigates how Monte Carlo (MC) and randomized Quasi-Monte Carlo (QMC) algorithms can efficiently identify feasible portfolios while assessing numerical accuracy\, computational cost\, and sensitivity to market conditions. \n \n \n \n \n \n \n \n \n \n \n Solution Researchers at the Institute of Information and Communication Technologies of the Bulgarian Academy of Sciences (IICT-BAS) developed a reproducible Python-based framework for stochastic portfolio optimization using the HEMUS supercomputer. The methodology compares classical Monte Carlo sampling with Mersenne Twister and randomized Quasi-Monte Carlo methods based on scrambled Sobol and randomized Halton sequences. Candidate portfolios are generated under full-investment and long-only constraints\, with a maximum allocation of 30% per asset. We investigate two optimization objectives: maximizing the Sharpe ratio to improve risk-adjusted returns and minimizing historical CVaR to limit losses in the worst 5% of monthly observations. The analysis considers three historical estimation windows of 3\, 5\, and 10 years and five candidate budgets ranging from 128 to 2048. We evaluate each configuration through 50 independent randomized repetitions and compare it with deterministic numerical benchmarks using approximation errors\, portfolio characteristics\, and execution times. This computational framework enables systematic analysis of how sampling techniques\, risk objectives and historical periods influence portfolio selection. Scientific impact Methodology and Computational Framework The research presents a reproducible Python-based framework for stochastic portfolio optimization\, implemented on the HEMUS supercomputer at IICT-BAS. The study examines 32 global financial assets across 3-\, 5- and 10-year historical windows\, comparing Monte Carlo (with pseudorandom generator Mersenne Twister) and randomized Quasi-Monte Carlo (with scrambled Sobol and Halton) methods under Sharpe ratio maximization and CVaR minimization. Experiment 1 – Accuracy and Convergence of Stochastic Portfolio Search: Five sampling budgets (128–2048 candidate portfolios) and 50 independent repetitions per configuration were evaluated against deterministic numerical benchmarks. Increasing the number of candidates reduced approximation errors\, while differences between sampling methods remained modest\, with classical Monte Carlo generally requiring less execution time.Experiment 2 – Financial Risk – Return Comparison: The 5-year Sharpe-oriented portfolios achieved a mean annualized return of 9.15%\, while the 10-year CVaR-oriented portfolios recorded the lowest mean historical monthly CVaR of 4.40% and annualized volatility of 6.59%. These results highlight the trade-off between maximizing risk-adjusted returns and minimizing downside risk.Experiment 3 – Historical Window and Sampling Sensitivity: The 3-\, 5- and 10-year estimation windows produced different portfolio allocations and risk–return characteristics. The optimization objective and historical period generally had a greater influence on portfolio selection than the choice of sampling generator.Scientific Contribution The study demonstrates the potential of HPC-supported stochastic optimization for reproducible financial risk analysis and numerical benchmarking. The HEMUS supercomputer enabled extensive computational experiments\, providing a foundation for scalable portfolio optimization and future financial decision-support systems. Benefits:  Improved optimization accuracy: Larger sampling budgets reduce approximation errors relative to deterministic benchmarks.Risk-aware portfolio selection: Compares return-oriented Sharpe optimization with downside-risk-focused CVaR.Reliable numerical benchmarking: Evaluates accuracy\, stability and computational efficiency across MC/RQMC methods.Better understanding of market sensitivity: Reveals how historical estimation windows influence investment outcomes.HPC-enabled research: Supports repeated simulations and scalable numerical experimentation.Future financial applications: Provides a foundation for computational portfolio analysis and investment decision-support systems.\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Success story # Highlights: Keywords: HPC; Monte Carlo; Quasi-Monte Carlo; Portfolio Optimization; Sharpe Ratio; CVaR; Financial Risk.Technology: HEMUS Supercomputer; Python; Mersenne Twister; Scrambled Sobol and Halton; Stochastic Optimization.Industry sectors: Finance/Insurance; IT/HPC Systems\, Services & Software Providers.Scientific achievement: Development of a reproducible HPC-based framework for comparing stochastic portfolio optimization methods.Key results: Larger sampling budgets improved numerical accuracy\, while optimization objectives and historical windows significantly influenced portfolio risk–return characteristics.HPC contribution: HEMUS enabled repeated simulations\, algorithm benchmarking and computational performance analysis.\n \n \n \n \n Comparison of the 3 time horizons through Sharpe-ratio optimization. Among the sampled candidates\, the 5-year Sharpe solution provides the highest expected return and the strongest reward-to-risk trade-off\, at the cost of the highest volatility across the three horizons \n \n \n \n \n Figure 1. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 2. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 3. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Comparison of the 3 time horizons through CVaR (5%). At N=100\, Mean-CVaR shifts the selected portfolios toward lower-volatility regions. The 5-year solution achieves the highest expected return while slightly reducing volatility relative to the 3-year case\, whereas the 10-year solution is the most conservative in volatility terms. \n \n \n \n \n Figure 4. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 5. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 6. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 7. Sharpe vs CVaR: Cross-Horizon Comparison. Sharpe is higher at all three horizons. CVaR is lower at all three horizons. \n \n \n \n \n \n \n \n \n \n \n Contact: Silvi-Maria Gurova | IICT-BAS | silvimaria.gurova@iict.bas.bgTodor Gurov | IICT-BAS | todor.gurov@iict.bas.bgAleksandar Kirilov | IICT-BAS | alex.kirilov@iict.bas.bg URL:https://eurocc-bulgaria.bg/?tribe_events=doctoral-information-meeting-on-february-5-2026 END:VEVENT BEGIN:VEVENT DTSTART;TZID=Europe/Helsinki:20260129T100000 DTEND;TZID=Europe/Helsinki:20260129T120000 DTSTAMP:20261023T101421 CREATED:20260119T080029Z LAST-MODIFIED:20260319T083043Z UID:7110-1769680800-1769688000@eurocc-bulgaria.bg SUMMARY:“Gray Scott Thursdays” webinar series 2026 kick-off webinar\, 29 January 2026\, 10:00AM\, online DESCRIPTION:NCC BULGARIA\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n NCC presenting the success story: NCC Bulgaria is the National Competence Centre for HPC and HPDA\, operating within the EuroCC3 network and EuroHPC ecosystem. Coordinated by the Institute of Information and Communication Technologies at the Bulgarian Academy of Sciences (IICT-BAS) in partnership with the Sofia University “St. Kliment Ohridski” (SU) and the University of National and World Economy (UNWE)\, it supports industry\, academia\, and public organizations through specialized training\, consulting\, technical expertise\, and access to national and European HPC resources. Scientific partners involved The Institute of Information and Communication Technologies (IICT-BAS) conducts fundamental and applied scientific research in the fields of high-performance computing (HPC)\, machine learning\, big data processing\, cybersecurity\, optimization\, financial modeling and mathematical modeling. The developed methods and technologies find application in medicine\, ecology\, industry\, transport\, financial analysis\, process optimization\, and digital transformation. Technical/scientific Challenge Constructing reliable investment portfolios requires balancing expected returns\, market volatility\, and potential losses under adverse market conditions. For a diversified universe of 32 global financial assets\, the large number of possible allocations makes constrained portfolio optimization computationally demanding. An additional challenge is that different risk measures\, such as the Sharpe ratio and Conditional Value-at-Risk (CVaR)\, may favour substantially different investment strategies. Portfolio decisions are also sensitive to the historical estimation period and the numerical sampling method. The research therefore investigates how Monte Carlo (MC) and randomized Quasi-Monte Carlo (QMC) algorithms can efficiently identify feasible portfolios while assessing numerical accuracy\, computational cost\, and sensitivity to market conditions. \n \n \n \n \n \n \n \n \n \n \n Solution Researchers at the Institute of Information and Communication Technologies of the Bulgarian Academy of Sciences (IICT-BAS) developed a reproducible Python-based framework for stochastic portfolio optimization using the HEMUS supercomputer. The methodology compares classical Monte Carlo sampling with Mersenne Twister and randomized Quasi-Monte Carlo methods based on scrambled Sobol and randomized Halton sequences. Candidate portfolios are generated under full-investment and long-only constraints\, with a maximum allocation of 30% per asset. We investigate two optimization objectives: maximizing the Sharpe ratio to improve risk-adjusted returns and minimizing historical CVaR to limit losses in the worst 5% of monthly observations. The analysis considers three historical estimation windows of 3\, 5\, and 10 years and five candidate budgets ranging from 128 to 2048. We evaluate each configuration through 50 independent randomized repetitions and compare it with deterministic numerical benchmarks using approximation errors\, portfolio characteristics\, and execution times. This computational framework enables systematic analysis of how sampling techniques\, risk objectives and historical periods influence portfolio selection. Scientific impact Methodology and Computational Framework The research presents a reproducible Python-based framework for stochastic portfolio optimization\, implemented on the HEMUS supercomputer at IICT-BAS. The study examines 32 global financial assets across 3-\, 5- and 10-year historical windows\, comparing Monte Carlo (with pseudorandom generator Mersenne Twister) and randomized Quasi-Monte Carlo (with scrambled Sobol and Halton) methods under Sharpe ratio maximization and CVaR minimization. Experiment 1 – Accuracy and Convergence of Stochastic Portfolio Search: Five sampling budgets (128–2048 candidate portfolios) and 50 independent repetitions per configuration were evaluated against deterministic numerical benchmarks. Increasing the number of candidates reduced approximation errors\, while differences between sampling methods remained modest\, with classical Monte Carlo generally requiring less execution time.Experiment 2 – Financial Risk – Return Comparison: The 5-year Sharpe-oriented portfolios achieved a mean annualized return of 9.15%\, while the 10-year CVaR-oriented portfolios recorded the lowest mean historical monthly CVaR of 4.40% and annualized volatility of 6.59%. These results highlight the trade-off between maximizing risk-adjusted returns and minimizing downside risk.Experiment 3 – Historical Window and Sampling Sensitivity: The 3-\, 5- and 10-year estimation windows produced different portfolio allocations and risk–return characteristics. The optimization objective and historical period generally had a greater influence on portfolio selection than the choice of sampling generator.Scientific Contribution The study demonstrates the potential of HPC-supported stochastic optimization for reproducible financial risk analysis and numerical benchmarking. The HEMUS supercomputer enabled extensive computational experiments\, providing a foundation for scalable portfolio optimization and future financial decision-support systems. Benefits:  Improved optimization accuracy: Larger sampling budgets reduce approximation errors relative to deterministic benchmarks.Risk-aware portfolio selection: Compares return-oriented Sharpe optimization with downside-risk-focused CVaR.Reliable numerical benchmarking: Evaluates accuracy\, stability and computational efficiency across MC/RQMC methods.Better understanding of market sensitivity: Reveals how historical estimation windows influence investment outcomes.HPC-enabled research: Supports repeated simulations and scalable numerical experimentation.Future financial applications: Provides a foundation for computational portfolio analysis and investment decision-support systems.\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Success story # Highlights: Keywords: HPC; Monte Carlo; Quasi-Monte Carlo; Portfolio Optimization; Sharpe Ratio; CVaR; Financial Risk.Technology: HEMUS Supercomputer; Python; Mersenne Twister; Scrambled Sobol and Halton; Stochastic Optimization.Industry sectors: Finance/Insurance; IT/HPC Systems\, Services & Software Providers.Scientific achievement: Development of a reproducible HPC-based framework for comparing stochastic portfolio optimization methods.Key results: Larger sampling budgets improved numerical accuracy\, while optimization objectives and historical windows significantly influenced portfolio risk–return characteristics.HPC contribution: HEMUS enabled repeated simulations\, algorithm benchmarking and computational performance analysis.\n \n \n \n \n Comparison of the 3 time horizons through Sharpe-ratio optimization. Among the sampled candidates\, the 5-year Sharpe solution provides the highest expected return and the strongest reward-to-risk trade-off\, at the cost of the highest volatility across the three horizons \n \n \n \n \n Figure 1. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 2. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 3. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Comparison of the 3 time horizons through CVaR (5%). At N=100\, Mean-CVaR shifts the selected portfolios toward lower-volatility regions. The 5-year solution achieves the highest expected return while slightly reducing volatility relative to the 3-year case\, whereas the 10-year solution is the most conservative in volatility terms. \n \n \n \n \n Figure 4. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 5. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 6. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 7. Sharpe vs CVaR: Cross-Horizon Comparison. Sharpe is higher at all three horizons. CVaR is lower at all three horizons. \n \n \n \n \n \n \n \n \n \n \n Contact: Silvi-Maria Gurova | IICT-BAS | silvimaria.gurova@iict.bas.bgTodor Gurov | IICT-BAS | todor.gurov@iict.bas.bgAleksandar Kirilov | IICT-BAS | alex.kirilov@iict.bas.bg URL:https://eurocc-bulgaria.bg/?tribe_events=gray-scott-thursdays-webinar-series-2026-kick-off-webinar-29-january-2026-1000am-online END:VEVENT BEGIN:VEVENT DTSTART;VALUE=DATE:20260112 DTEND;VALUE=DATE:20260117 DTSTAMP:20261023T101421 CREATED:20260102T080043Z LAST-MODIFIED:20260319T082728Z UID:7105-1768176000-1768607999@eurocc-bulgaria.bg SUMMARY:GENE/GENE-X user training and tutorial\, organized by CoE Plasma-PEPSC\, ENCCS\, NCC Germany\, 12-16 January\, hybrid DESCRIPTION:NCC BULGARIA\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n NCC presenting the success story: NCC Bulgaria is the National Competence Centre for HPC and HPDA\, operating within the EuroCC3 network and EuroHPC ecosystem. Coordinated by the Institute of Information and Communication Technologies at the Bulgarian Academy of Sciences (IICT-BAS) in partnership with the Sofia University “St. Kliment Ohridski” (SU) and the University of National and World Economy (UNWE)\, it supports industry\, academia\, and public organizations through specialized training\, consulting\, technical expertise\, and access to national and European HPC resources. Scientific partners involved The Institute of Information and Communication Technologies (IICT-BAS) conducts fundamental and applied scientific research in the fields of high-performance computing (HPC)\, machine learning\, big data processing\, cybersecurity\, optimization\, financial modeling and mathematical modeling. The developed methods and technologies find application in medicine\, ecology\, industry\, transport\, financial analysis\, process optimization\, and digital transformation. Technical/scientific Challenge Constructing reliable investment portfolios requires balancing expected returns\, market volatility\, and potential losses under adverse market conditions. For a diversified universe of 32 global financial assets\, the large number of possible allocations makes constrained portfolio optimization computationally demanding. An additional challenge is that different risk measures\, such as the Sharpe ratio and Conditional Value-at-Risk (CVaR)\, may favour substantially different investment strategies. Portfolio decisions are also sensitive to the historical estimation period and the numerical sampling method. The research therefore investigates how Monte Carlo (MC) and randomized Quasi-Monte Carlo (QMC) algorithms can efficiently identify feasible portfolios while assessing numerical accuracy\, computational cost\, and sensitivity to market conditions. \n \n \n \n \n \n \n \n \n \n \n Solution Researchers at the Institute of Information and Communication Technologies of the Bulgarian Academy of Sciences (IICT-BAS) developed a reproducible Python-based framework for stochastic portfolio optimization using the HEMUS supercomputer. The methodology compares classical Monte Carlo sampling with Mersenne Twister and randomized Quasi-Monte Carlo methods based on scrambled Sobol and randomized Halton sequences. Candidate portfolios are generated under full-investment and long-only constraints\, with a maximum allocation of 30% per asset. We investigate two optimization objectives: maximizing the Sharpe ratio to improve risk-adjusted returns and minimizing historical CVaR to limit losses in the worst 5% of monthly observations. The analysis considers three historical estimation windows of 3\, 5\, and 10 years and five candidate budgets ranging from 128 to 2048. We evaluate each configuration through 50 independent randomized repetitions and compare it with deterministic numerical benchmarks using approximation errors\, portfolio characteristics\, and execution times. This computational framework enables systematic analysis of how sampling techniques\, risk objectives and historical periods influence portfolio selection. Scientific impact Methodology and Computational Framework The research presents a reproducible Python-based framework for stochastic portfolio optimization\, implemented on the HEMUS supercomputer at IICT-BAS. The study examines 32 global financial assets across 3-\, 5- and 10-year historical windows\, comparing Monte Carlo (with pseudorandom generator Mersenne Twister) and randomized Quasi-Monte Carlo (with scrambled Sobol and Halton) methods under Sharpe ratio maximization and CVaR minimization. Experiment 1 – Accuracy and Convergence of Stochastic Portfolio Search: Five sampling budgets (128–2048 candidate portfolios) and 50 independent repetitions per configuration were evaluated against deterministic numerical benchmarks. Increasing the number of candidates reduced approximation errors\, while differences between sampling methods remained modest\, with classical Monte Carlo generally requiring less execution time.Experiment 2 – Financial Risk – Return Comparison: The 5-year Sharpe-oriented portfolios achieved a mean annualized return of 9.15%\, while the 10-year CVaR-oriented portfolios recorded the lowest mean historical monthly CVaR of 4.40% and annualized volatility of 6.59%. These results highlight the trade-off between maximizing risk-adjusted returns and minimizing downside risk.Experiment 3 – Historical Window and Sampling Sensitivity: The 3-\, 5- and 10-year estimation windows produced different portfolio allocations and risk–return characteristics. The optimization objective and historical period generally had a greater influence on portfolio selection than the choice of sampling generator.Scientific Contribution The study demonstrates the potential of HPC-supported stochastic optimization for reproducible financial risk analysis and numerical benchmarking. The HEMUS supercomputer enabled extensive computational experiments\, providing a foundation for scalable portfolio optimization and future financial decision-support systems. Benefits:  Improved optimization accuracy: Larger sampling budgets reduce approximation errors relative to deterministic benchmarks.Risk-aware portfolio selection: Compares return-oriented Sharpe optimization with downside-risk-focused CVaR.Reliable numerical benchmarking: Evaluates accuracy\, stability and computational efficiency across MC/RQMC methods.Better understanding of market sensitivity: Reveals how historical estimation windows influence investment outcomes.HPC-enabled research: Supports repeated simulations and scalable numerical experimentation.Future financial applications: Provides a foundation for computational portfolio analysis and investment decision-support systems.\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Success story # Highlights: Keywords: HPC; Monte Carlo; Quasi-Monte Carlo; Portfolio Optimization; Sharpe Ratio; CVaR; Financial Risk.Technology: HEMUS Supercomputer; Python; Mersenne Twister; Scrambled Sobol and Halton; Stochastic Optimization.Industry sectors: Finance/Insurance; IT/HPC Systems\, Services & Software Providers.Scientific achievement: Development of a reproducible HPC-based framework for comparing stochastic portfolio optimization methods.Key results: Larger sampling budgets improved numerical accuracy\, while optimization objectives and historical windows significantly influenced portfolio risk–return characteristics.HPC contribution: HEMUS enabled repeated simulations\, algorithm benchmarking and computational performance analysis.\n \n \n \n \n Comparison of the 3 time horizons through Sharpe-ratio optimization. Among the sampled candidates\, the 5-year Sharpe solution provides the highest expected return and the strongest reward-to-risk trade-off\, at the cost of the highest volatility across the three horizons \n \n \n \n \n Figure 1. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 2. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 3. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Comparison of the 3 time horizons through CVaR (5%). At N=100\, Mean-CVaR shifts the selected portfolios toward lower-volatility regions. The 5-year solution achieves the highest expected return while slightly reducing volatility relative to the 3-year case\, whereas the 10-year solution is the most conservative in volatility terms. \n \n \n \n \n Figure 4. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 5. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 6. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 7. Sharpe vs CVaR: Cross-Horizon Comparison. Sharpe is higher at all three horizons. CVaR is lower at all three horizons. \n \n \n \n \n \n \n \n \n \n \n Contact: Silvi-Maria Gurova | IICT-BAS | silvimaria.gurova@iict.bas.bgTodor Gurov | IICT-BAS | todor.gurov@iict.bas.bgAleksandar Kirilov | IICT-BAS | alex.kirilov@iict.bas.bg URL:https://eurocc-bulgaria.bg/?tribe_events=gene-gene-x-user-training-and-tutorial-organized-by-coe-plasma-pepsc-enccs-ncc-germany-12-16-january-hybrid ATTACH;FMTTYPE=image/png:https://eurocc-bulgaria.bg/wp-content/uploads/2026/03/GENE_GENE-X_Garching-Munich_workshop_2025_zoomout6_1920x1080.png END:VEVENT BEGIN:VEVENT DTSTART;TZID=Europe/Helsinki:20251215T100000 DTEND;TZID=Europe/Helsinki:20251215T130000 DTSTAMP:20261023T101421 CREATED:20251212T133353Z LAST-MODIFIED:20251212T133353Z UID:6769-1765792800-1765803600@eurocc-bulgaria.bg SUMMARY:Training: Development of HPC Applications\, 15 December 2025\, Sofia DESCRIPTION:NCC BULGARIA\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n NCC presenting the success story: NCC Bulgaria is the National Competence Centre for HPC and HPDA\, operating within the EuroCC3 network and EuroHPC ecosystem. Coordinated by the Institute of Information and Communication Technologies at the Bulgarian Academy of Sciences (IICT-BAS) in partnership with the Sofia University “St. Kliment Ohridski” (SU) and the University of National and World Economy (UNWE)\, it supports industry\, academia\, and public organizations through specialized training\, consulting\, technical expertise\, and access to national and European HPC resources. Scientific partners involved The Institute of Information and Communication Technologies (IICT-BAS) conducts fundamental and applied scientific research in the fields of high-performance computing (HPC)\, machine learning\, big data processing\, cybersecurity\, optimization\, financial modeling and mathematical modeling. The developed methods and technologies find application in medicine\, ecology\, industry\, transport\, financial analysis\, process optimization\, and digital transformation. Technical/scientific Challenge Constructing reliable investment portfolios requires balancing expected returns\, market volatility\, and potential losses under adverse market conditions. For a diversified universe of 32 global financial assets\, the large number of possible allocations makes constrained portfolio optimization computationally demanding. An additional challenge is that different risk measures\, such as the Sharpe ratio and Conditional Value-at-Risk (CVaR)\, may favour substantially different investment strategies. Portfolio decisions are also sensitive to the historical estimation period and the numerical sampling method. The research therefore investigates how Monte Carlo (MC) and randomized Quasi-Monte Carlo (QMC) algorithms can efficiently identify feasible portfolios while assessing numerical accuracy\, computational cost\, and sensitivity to market conditions. \n \n \n \n \n \n \n \n \n \n \n Solution Researchers at the Institute of Information and Communication Technologies of the Bulgarian Academy of Sciences (IICT-BAS) developed a reproducible Python-based framework for stochastic portfolio optimization using the HEMUS supercomputer. The methodology compares classical Monte Carlo sampling with Mersenne Twister and randomized Quasi-Monte Carlo methods based on scrambled Sobol and randomized Halton sequences. Candidate portfolios are generated under full-investment and long-only constraints\, with a maximum allocation of 30% per asset. We investigate two optimization objectives: maximizing the Sharpe ratio to improve risk-adjusted returns and minimizing historical CVaR to limit losses in the worst 5% of monthly observations. The analysis considers three historical estimation windows of 3\, 5\, and 10 years and five candidate budgets ranging from 128 to 2048. We evaluate each configuration through 50 independent randomized repetitions and compare it with deterministic numerical benchmarks using approximation errors\, portfolio characteristics\, and execution times. This computational framework enables systematic analysis of how sampling techniques\, risk objectives and historical periods influence portfolio selection. Scientific impact Methodology and Computational Framework The research presents a reproducible Python-based framework for stochastic portfolio optimization\, implemented on the HEMUS supercomputer at IICT-BAS. The study examines 32 global financial assets across 3-\, 5- and 10-year historical windows\, comparing Monte Carlo (with pseudorandom generator Mersenne Twister) and randomized Quasi-Monte Carlo (with scrambled Sobol and Halton) methods under Sharpe ratio maximization and CVaR minimization. Experiment 1 – Accuracy and Convergence of Stochastic Portfolio Search: Five sampling budgets (128–2048 candidate portfolios) and 50 independent repetitions per configuration were evaluated against deterministic numerical benchmarks. Increasing the number of candidates reduced approximation errors\, while differences between sampling methods remained modest\, with classical Monte Carlo generally requiring less execution time.Experiment 2 – Financial Risk – Return Comparison: The 5-year Sharpe-oriented portfolios achieved a mean annualized return of 9.15%\, while the 10-year CVaR-oriented portfolios recorded the lowest mean historical monthly CVaR of 4.40% and annualized volatility of 6.59%. These results highlight the trade-off between maximizing risk-adjusted returns and minimizing downside risk.Experiment 3 – Historical Window and Sampling Sensitivity: The 3-\, 5- and 10-year estimation windows produced different portfolio allocations and risk–return characteristics. The optimization objective and historical period generally had a greater influence on portfolio selection than the choice of sampling generator.Scientific Contribution The study demonstrates the potential of HPC-supported stochastic optimization for reproducible financial risk analysis and numerical benchmarking. The HEMUS supercomputer enabled extensive computational experiments\, providing a foundation for scalable portfolio optimization and future financial decision-support systems. Benefits:  Improved optimization accuracy: Larger sampling budgets reduce approximation errors relative to deterministic benchmarks.Risk-aware portfolio selection: Compares return-oriented Sharpe optimization with downside-risk-focused CVaR.Reliable numerical benchmarking: Evaluates accuracy\, stability and computational efficiency across MC/RQMC methods.Better understanding of market sensitivity: Reveals how historical estimation windows influence investment outcomes.HPC-enabled research: Supports repeated simulations and scalable numerical experimentation.Future financial applications: Provides a foundation for computational portfolio analysis and investment decision-support systems.\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Success story # Highlights: Keywords: HPC; Monte Carlo; Quasi-Monte Carlo; Portfolio Optimization; Sharpe Ratio; CVaR; Financial Risk.Technology: HEMUS Supercomputer; Python; Mersenne Twister; Scrambled Sobol and Halton; Stochastic Optimization.Industry sectors: Finance/Insurance; IT/HPC Systems\, Services & Software Providers.Scientific achievement: Development of a reproducible HPC-based framework for comparing stochastic portfolio optimization methods.Key results: Larger sampling budgets improved numerical accuracy\, while optimization objectives and historical windows significantly influenced portfolio risk–return characteristics.HPC contribution: HEMUS enabled repeated simulations\, algorithm benchmarking and computational performance analysis.\n \n \n \n \n Comparison of the 3 time horizons through Sharpe-ratio optimization. Among the sampled candidates\, the 5-year Sharpe solution provides the highest expected return and the strongest reward-to-risk trade-off\, at the cost of the highest volatility across the three horizons \n \n \n \n \n Figure 1. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 2. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 3. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Comparison of the 3 time horizons through CVaR (5%). At N=100\, Mean-CVaR shifts the selected portfolios toward lower-volatility regions. The 5-year solution achieves the highest expected return while slightly reducing volatility relative to the 3-year case\, whereas the 10-year solution is the most conservative in volatility terms. \n \n \n \n \n Figure 4. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 5. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 6. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 7. Sharpe vs CVaR: Cross-Horizon Comparison. Sharpe is higher at all three horizons. CVaR is lower at all three horizons. \n \n \n \n \n \n \n \n \n \n \n Contact: Silvi-Maria Gurova | IICT-BAS | silvimaria.gurova@iict.bas.bgTodor Gurov | IICT-BAS | todor.gurov@iict.bas.bgAleksandar Kirilov | IICT-BAS | alex.kirilov@iict.bas.bg URL:https://eurocc-bulgaria.bg/?tribe_events=training-development-of-hpc-applications-15-december-2025-sofia ATTACH;FMTTYPE=image/jpeg:https://eurocc-bulgaria.bg/wp-content/uploads/2025/12/HPC20Training20220112008_05_2420PM.jpg END:VEVENT BEGIN:VEVENT DTSTART;TZID=Europe/Helsinki:20251210T080000 DTEND;TZID=Europe/Helsinki:20251210T170000 DTSTAMP:20261023T101421 CREATED:20251125T130335Z LAST-MODIFIED:20251125T130335Z UID:6679-1765353600-1765386000@eurocc-bulgaria.bg SUMMARY:20th Annual Meeting of the Bulgarian Section of SIAM\, December 10-12\, 2025 DESCRIPTION:NCC BULGARIA\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n NCC presenting the success story: NCC Bulgaria is the National Competence Centre for HPC and HPDA\, operating within the EuroCC3 network and EuroHPC ecosystem. Coordinated by the Institute of Information and Communication Technologies at the Bulgarian Academy of Sciences (IICT-BAS) in partnership with the Sofia University “St. Kliment Ohridski” (SU) and the University of National and World Economy (UNWE)\, it supports industry\, academia\, and public organizations through specialized training\, consulting\, technical expertise\, and access to national and European HPC resources. Scientific partners involved The Institute of Information and Communication Technologies (IICT-BAS) conducts fundamental and applied scientific research in the fields of high-performance computing (HPC)\, machine learning\, big data processing\, cybersecurity\, optimization\, financial modeling and mathematical modeling. The developed methods and technologies find application in medicine\, ecology\, industry\, transport\, financial analysis\, process optimization\, and digital transformation. Technical/scientific Challenge Constructing reliable investment portfolios requires balancing expected returns\, market volatility\, and potential losses under adverse market conditions. For a diversified universe of 32 global financial assets\, the large number of possible allocations makes constrained portfolio optimization computationally demanding. An additional challenge is that different risk measures\, such as the Sharpe ratio and Conditional Value-at-Risk (CVaR)\, may favour substantially different investment strategies. Portfolio decisions are also sensitive to the historical estimation period and the numerical sampling method. The research therefore investigates how Monte Carlo (MC) and randomized Quasi-Monte Carlo (QMC) algorithms can efficiently identify feasible portfolios while assessing numerical accuracy\, computational cost\, and sensitivity to market conditions. \n \n \n \n \n \n \n \n \n \n \n Solution Researchers at the Institute of Information and Communication Technologies of the Bulgarian Academy of Sciences (IICT-BAS) developed a reproducible Python-based framework for stochastic portfolio optimization using the HEMUS supercomputer. The methodology compares classical Monte Carlo sampling with Mersenne Twister and randomized Quasi-Monte Carlo methods based on scrambled Sobol and randomized Halton sequences. Candidate portfolios are generated under full-investment and long-only constraints\, with a maximum allocation of 30% per asset. We investigate two optimization objectives: maximizing the Sharpe ratio to improve risk-adjusted returns and minimizing historical CVaR to limit losses in the worst 5% of monthly observations. The analysis considers three historical estimation windows of 3\, 5\, and 10 years and five candidate budgets ranging from 128 to 2048. We evaluate each configuration through 50 independent randomized repetitions and compare it with deterministic numerical benchmarks using approximation errors\, portfolio characteristics\, and execution times. This computational framework enables systematic analysis of how sampling techniques\, risk objectives and historical periods influence portfolio selection. Scientific impact Methodology and Computational Framework The research presents a reproducible Python-based framework for stochastic portfolio optimization\, implemented on the HEMUS supercomputer at IICT-BAS. The study examines 32 global financial assets across 3-\, 5- and 10-year historical windows\, comparing Monte Carlo (with pseudorandom generator Mersenne Twister) and randomized Quasi-Monte Carlo (with scrambled Sobol and Halton) methods under Sharpe ratio maximization and CVaR minimization. Experiment 1 – Accuracy and Convergence of Stochastic Portfolio Search: Five sampling budgets (128–2048 candidate portfolios) and 50 independent repetitions per configuration were evaluated against deterministic numerical benchmarks. Increasing the number of candidates reduced approximation errors\, while differences between sampling methods remained modest\, with classical Monte Carlo generally requiring less execution time.Experiment 2 – Financial Risk – Return Comparison: The 5-year Sharpe-oriented portfolios achieved a mean annualized return of 9.15%\, while the 10-year CVaR-oriented portfolios recorded the lowest mean historical monthly CVaR of 4.40% and annualized volatility of 6.59%. These results highlight the trade-off between maximizing risk-adjusted returns and minimizing downside risk.Experiment 3 – Historical Window and Sampling Sensitivity: The 3-\, 5- and 10-year estimation windows produced different portfolio allocations and risk–return characteristics. The optimization objective and historical period generally had a greater influence on portfolio selection than the choice of sampling generator.Scientific Contribution The study demonstrates the potential of HPC-supported stochastic optimization for reproducible financial risk analysis and numerical benchmarking. The HEMUS supercomputer enabled extensive computational experiments\, providing a foundation for scalable portfolio optimization and future financial decision-support systems. Benefits:  Improved optimization accuracy: Larger sampling budgets reduce approximation errors relative to deterministic benchmarks.Risk-aware portfolio selection: Compares return-oriented Sharpe optimization with downside-risk-focused CVaR.Reliable numerical benchmarking: Evaluates accuracy\, stability and computational efficiency across MC/RQMC methods.Better understanding of market sensitivity: Reveals how historical estimation windows influence investment outcomes.HPC-enabled research: Supports repeated simulations and scalable numerical experimentation.Future financial applications: Provides a foundation for computational portfolio analysis and investment decision-support systems.\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Success story # Highlights: Keywords: HPC; Monte Carlo; Quasi-Monte Carlo; Portfolio Optimization; Sharpe Ratio; CVaR; Financial Risk.Technology: HEMUS Supercomputer; Python; Mersenne Twister; Scrambled Sobol and Halton; Stochastic Optimization.Industry sectors: Finance/Insurance; IT/HPC Systems\, Services & Software Providers.Scientific achievement: Development of a reproducible HPC-based framework for comparing stochastic portfolio optimization methods.Key results: Larger sampling budgets improved numerical accuracy\, while optimization objectives and historical windows significantly influenced portfolio risk–return characteristics.HPC contribution: HEMUS enabled repeated simulations\, algorithm benchmarking and computational performance analysis.\n \n \n \n \n Comparison of the 3 time horizons through Sharpe-ratio optimization. Among the sampled candidates\, the 5-year Sharpe solution provides the highest expected return and the strongest reward-to-risk trade-off\, at the cost of the highest volatility across the three horizons \n \n \n \n \n Figure 1. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 2. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 3. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Comparison of the 3 time horizons through CVaR (5%). At N=100\, Mean-CVaR shifts the selected portfolios toward lower-volatility regions. The 5-year solution achieves the highest expected return while slightly reducing volatility relative to the 3-year case\, whereas the 10-year solution is the most conservative in volatility terms. \n \n \n \n \n Figure 4. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 5. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 6. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 7. Sharpe vs CVaR: Cross-Horizon Comparison. Sharpe is higher at all three horizons. CVaR is lower at all three horizons. \n \n \n \n \n \n \n \n \n \n \n Contact: Silvi-Maria Gurova | IICT-BAS | silvimaria.gurova@iict.bas.bgTodor Gurov | IICT-BAS | todor.gurov@iict.bas.bgAleksandar Kirilov | IICT-BAS | alex.kirilov@iict.bas.bg URL:https://eurocc-bulgaria.bg/?tribe_events=20th-annual-meeting-of-the-bulgarian-section-of-siam-december-10-12-2025 END:VEVENT BEGIN:VEVENT DTSTART;TZID=Europe/Helsinki:20251208T100000 DTEND;TZID=Europe/Helsinki:20251208T130000 DTSTAMP:20261023T101421 CREATED:20251201T080011Z LAST-MODIFIED:20260319T082430Z UID:7100-1765188000-1765198800@eurocc-bulgaria.bg SUMMARY:Containerisation workshop\, NCC Hungary\, 8 December 2025 09:00-12:00\, online DESCRIPTION:NCC BULGARIA\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n NCC presenting the success story: NCC Bulgaria is the National Competence Centre for HPC and HPDA\, operating within the EuroCC3 network and EuroHPC ecosystem. Coordinated by the Institute of Information and Communication Technologies at the Bulgarian Academy of Sciences (IICT-BAS) in partnership with the Sofia University “St. Kliment Ohridski” (SU) and the University of National and World Economy (UNWE)\, it supports industry\, academia\, and public organizations through specialized training\, consulting\, technical expertise\, and access to national and European HPC resources. Scientific partners involved The Institute of Information and Communication Technologies (IICT-BAS) conducts fundamental and applied scientific research in the fields of high-performance computing (HPC)\, machine learning\, big data processing\, cybersecurity\, optimization\, financial modeling and mathematical modeling. The developed methods and technologies find application in medicine\, ecology\, industry\, transport\, financial analysis\, process optimization\, and digital transformation. Technical/scientific Challenge Constructing reliable investment portfolios requires balancing expected returns\, market volatility\, and potential losses under adverse market conditions. For a diversified universe of 32 global financial assets\, the large number of possible allocations makes constrained portfolio optimization computationally demanding. An additional challenge is that different risk measures\, such as the Sharpe ratio and Conditional Value-at-Risk (CVaR)\, may favour substantially different investment strategies. Portfolio decisions are also sensitive to the historical estimation period and the numerical sampling method. The research therefore investigates how Monte Carlo (MC) and randomized Quasi-Monte Carlo (QMC) algorithms can efficiently identify feasible portfolios while assessing numerical accuracy\, computational cost\, and sensitivity to market conditions. \n \n \n \n \n \n \n \n \n \n \n Solution Researchers at the Institute of Information and Communication Technologies of the Bulgarian Academy of Sciences (IICT-BAS) developed a reproducible Python-based framework for stochastic portfolio optimization using the HEMUS supercomputer. The methodology compares classical Monte Carlo sampling with Mersenne Twister and randomized Quasi-Monte Carlo methods based on scrambled Sobol and randomized Halton sequences. Candidate portfolios are generated under full-investment and long-only constraints\, with a maximum allocation of 30% per asset. We investigate two optimization objectives: maximizing the Sharpe ratio to improve risk-adjusted returns and minimizing historical CVaR to limit losses in the worst 5% of monthly observations. The analysis considers three historical estimation windows of 3\, 5\, and 10 years and five candidate budgets ranging from 128 to 2048. We evaluate each configuration through 50 independent randomized repetitions and compare it with deterministic numerical benchmarks using approximation errors\, portfolio characteristics\, and execution times. This computational framework enables systematic analysis of how sampling techniques\, risk objectives and historical periods influence portfolio selection. Scientific impact Methodology and Computational Framework The research presents a reproducible Python-based framework for stochastic portfolio optimization\, implemented on the HEMUS supercomputer at IICT-BAS. The study examines 32 global financial assets across 3-\, 5- and 10-year historical windows\, comparing Monte Carlo (with pseudorandom generator Mersenne Twister) and randomized Quasi-Monte Carlo (with scrambled Sobol and Halton) methods under Sharpe ratio maximization and CVaR minimization. Experiment 1 – Accuracy and Convergence of Stochastic Portfolio Search: Five sampling budgets (128–2048 candidate portfolios) and 50 independent repetitions per configuration were evaluated against deterministic numerical benchmarks. Increasing the number of candidates reduced approximation errors\, while differences between sampling methods remained modest\, with classical Monte Carlo generally requiring less execution time.Experiment 2 – Financial Risk – Return Comparison: The 5-year Sharpe-oriented portfolios achieved a mean annualized return of 9.15%\, while the 10-year CVaR-oriented portfolios recorded the lowest mean historical monthly CVaR of 4.40% and annualized volatility of 6.59%. These results highlight the trade-off between maximizing risk-adjusted returns and minimizing downside risk.Experiment 3 – Historical Window and Sampling Sensitivity: The 3-\, 5- and 10-year estimation windows produced different portfolio allocations and risk–return characteristics. The optimization objective and historical period generally had a greater influence on portfolio selection than the choice of sampling generator.Scientific Contribution The study demonstrates the potential of HPC-supported stochastic optimization for reproducible financial risk analysis and numerical benchmarking. The HEMUS supercomputer enabled extensive computational experiments\, providing a foundation for scalable portfolio optimization and future financial decision-support systems. Benefits:  Improved optimization accuracy: Larger sampling budgets reduce approximation errors relative to deterministic benchmarks.Risk-aware portfolio selection: Compares return-oriented Sharpe optimization with downside-risk-focused CVaR.Reliable numerical benchmarking: Evaluates accuracy\, stability and computational efficiency across MC/RQMC methods.Better understanding of market sensitivity: Reveals how historical estimation windows influence investment outcomes.HPC-enabled research: Supports repeated simulations and scalable numerical experimentation.Future financial applications: Provides a foundation for computational portfolio analysis and investment decision-support systems.\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Success story # Highlights: Keywords: HPC; Monte Carlo; Quasi-Monte Carlo; Portfolio Optimization; Sharpe Ratio; CVaR; Financial Risk.Technology: HEMUS Supercomputer; Python; Mersenne Twister; Scrambled Sobol and Halton; Stochastic Optimization.Industry sectors: Finance/Insurance; IT/HPC Systems\, Services & Software Providers.Scientific achievement: Development of a reproducible HPC-based framework for comparing stochastic portfolio optimization methods.Key results: Larger sampling budgets improved numerical accuracy\, while optimization objectives and historical windows significantly influenced portfolio risk–return characteristics.HPC contribution: HEMUS enabled repeated simulations\, algorithm benchmarking and computational performance analysis.\n \n \n \n \n Comparison of the 3 time horizons through Sharpe-ratio optimization. Among the sampled candidates\, the 5-year Sharpe solution provides the highest expected return and the strongest reward-to-risk trade-off\, at the cost of the highest volatility across the three horizons \n \n \n \n \n Figure 1. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 2. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 3. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Comparison of the 3 time horizons through CVaR (5%). At N=100\, Mean-CVaR shifts the selected portfolios toward lower-volatility regions. The 5-year solution achieves the highest expected return while slightly reducing volatility relative to the 3-year case\, whereas the 10-year solution is the most conservative in volatility terms. \n \n \n \n \n Figure 4. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 5. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 6. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 7. Sharpe vs CVaR: Cross-Horizon Comparison. Sharpe is higher at all three horizons. CVaR is lower at all three horizons. \n \n \n \n \n \n \n \n \n \n \n Contact: Silvi-Maria Gurova | IICT-BAS | silvimaria.gurova@iict.bas.bgTodor Gurov | IICT-BAS | todor.gurov@iict.bas.bgAleksandar Kirilov | IICT-BAS | alex.kirilov@iict.bas.bg URL:https://eurocc-bulgaria.bg/?tribe_events=containerisation-workshop-ncc-hungary-8-december-2025-0900-1200-online ATTACH;FMTTYPE=image/png:https://eurocc-bulgaria.bg/wp-content/uploads/2026/03/HPC_workshop_banner_1126_kicsi_jav.png END:VEVENT BEGIN:VEVENT DTSTART;VALUE=DATE:20251126 DTEND;VALUE=DATE:20251129 DTSTAMP:20261023T101421 CREATED:20251110T080055Z LAST-MODIFIED:20260319T082031Z UID:7095-1764115200-1764374399@eurocc-bulgaria.bg SUMMARY:FROM BUILDING TO CITY ENERGY SIMULATION\, November 26 – 28\, online DESCRIPTION:NCC BULGARIA\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n NCC presenting the success story: NCC Bulgaria is the National Competence Centre for HPC and HPDA\, operating within the EuroCC3 network and EuroHPC ecosystem. Coordinated by the Institute of Information and Communication Technologies at the Bulgarian Academy of Sciences (IICT-BAS) in partnership with the Sofia University “St. Kliment Ohridski” (SU) and the University of National and World Economy (UNWE)\, it supports industry\, academia\, and public organizations through specialized training\, consulting\, technical expertise\, and access to national and European HPC resources. Scientific partners involved The Institute of Information and Communication Technologies (IICT-BAS) conducts fundamental and applied scientific research in the fields of high-performance computing (HPC)\, machine learning\, big data processing\, cybersecurity\, optimization\, financial modeling and mathematical modeling. The developed methods and technologies find application in medicine\, ecology\, industry\, transport\, financial analysis\, process optimization\, and digital transformation. Technical/scientific Challenge Constructing reliable investment portfolios requires balancing expected returns\, market volatility\, and potential losses under adverse market conditions. For a diversified universe of 32 global financial assets\, the large number of possible allocations makes constrained portfolio optimization computationally demanding. An additional challenge is that different risk measures\, such as the Sharpe ratio and Conditional Value-at-Risk (CVaR)\, may favour substantially different investment strategies. Portfolio decisions are also sensitive to the historical estimation period and the numerical sampling method. The research therefore investigates how Monte Carlo (MC) and randomized Quasi-Monte Carlo (QMC) algorithms can efficiently identify feasible portfolios while assessing numerical accuracy\, computational cost\, and sensitivity to market conditions. \n \n \n \n \n \n \n \n \n \n \n Solution Researchers at the Institute of Information and Communication Technologies of the Bulgarian Academy of Sciences (IICT-BAS) developed a reproducible Python-based framework for stochastic portfolio optimization using the HEMUS supercomputer. The methodology compares classical Monte Carlo sampling with Mersenne Twister and randomized Quasi-Monte Carlo methods based on scrambled Sobol and randomized Halton sequences. Candidate portfolios are generated under full-investment and long-only constraints\, with a maximum allocation of 30% per asset. We investigate two optimization objectives: maximizing the Sharpe ratio to improve risk-adjusted returns and minimizing historical CVaR to limit losses in the worst 5% of monthly observations. The analysis considers three historical estimation windows of 3\, 5\, and 10 years and five candidate budgets ranging from 128 to 2048. We evaluate each configuration through 50 independent randomized repetitions and compare it with deterministic numerical benchmarks using approximation errors\, portfolio characteristics\, and execution times. This computational framework enables systematic analysis of how sampling techniques\, risk objectives and historical periods influence portfolio selection. Scientific impact Methodology and Computational Framework The research presents a reproducible Python-based framework for stochastic portfolio optimization\, implemented on the HEMUS supercomputer at IICT-BAS. The study examines 32 global financial assets across 3-\, 5- and 10-year historical windows\, comparing Monte Carlo (with pseudorandom generator Mersenne Twister) and randomized Quasi-Monte Carlo (with scrambled Sobol and Halton) methods under Sharpe ratio maximization and CVaR minimization. Experiment 1 – Accuracy and Convergence of Stochastic Portfolio Search: Five sampling budgets (128–2048 candidate portfolios) and 50 independent repetitions per configuration were evaluated against deterministic numerical benchmarks. Increasing the number of candidates reduced approximation errors\, while differences between sampling methods remained modest\, with classical Monte Carlo generally requiring less execution time.Experiment 2 – Financial Risk – Return Comparison: The 5-year Sharpe-oriented portfolios achieved a mean annualized return of 9.15%\, while the 10-year CVaR-oriented portfolios recorded the lowest mean historical monthly CVaR of 4.40% and annualized volatility of 6.59%. These results highlight the trade-off between maximizing risk-adjusted returns and minimizing downside risk.Experiment 3 – Historical Window and Sampling Sensitivity: The 3-\, 5- and 10-year estimation windows produced different portfolio allocations and risk–return characteristics. The optimization objective and historical period generally had a greater influence on portfolio selection than the choice of sampling generator.Scientific Contribution The study demonstrates the potential of HPC-supported stochastic optimization for reproducible financial risk analysis and numerical benchmarking. The HEMUS supercomputer enabled extensive computational experiments\, providing a foundation for scalable portfolio optimization and future financial decision-support systems. Benefits:  Improved optimization accuracy: Larger sampling budgets reduce approximation errors relative to deterministic benchmarks.Risk-aware portfolio selection: Compares return-oriented Sharpe optimization with downside-risk-focused CVaR.Reliable numerical benchmarking: Evaluates accuracy\, stability and computational efficiency across MC/RQMC methods.Better understanding of market sensitivity: Reveals how historical estimation windows influence investment outcomes.HPC-enabled research: Supports repeated simulations and scalable numerical experimentation.Future financial applications: Provides a foundation for computational portfolio analysis and investment decision-support systems.\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Success story # Highlights: Keywords: HPC; Monte Carlo; Quasi-Monte Carlo; Portfolio Optimization; Sharpe Ratio; CVaR; Financial Risk.Technology: HEMUS Supercomputer; Python; Mersenne Twister; Scrambled Sobol and Halton; Stochastic Optimization.Industry sectors: Finance/Insurance; IT/HPC Systems\, Services & Software Providers.Scientific achievement: Development of a reproducible HPC-based framework for comparing stochastic portfolio optimization methods.Key results: Larger sampling budgets improved numerical accuracy\, while optimization objectives and historical windows significantly influenced portfolio risk–return characteristics.HPC contribution: HEMUS enabled repeated simulations\, algorithm benchmarking and computational performance analysis.\n \n \n \n \n Comparison of the 3 time horizons through Sharpe-ratio optimization. Among the sampled candidates\, the 5-year Sharpe solution provides the highest expected return and the strongest reward-to-risk trade-off\, at the cost of the highest volatility across the three horizons \n \n \n \n \n Figure 1. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 2. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 3. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Comparison of the 3 time horizons through CVaR (5%). At N=100\, Mean-CVaR shifts the selected portfolios toward lower-volatility regions. The 5-year solution achieves the highest expected return while slightly reducing volatility relative to the 3-year case\, whereas the 10-year solution is the most conservative in volatility terms. \n \n \n \n \n Figure 4. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 5. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 6. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 7. Sharpe vs CVaR: Cross-Horizon Comparison. Sharpe is higher at all three horizons. CVaR is lower at all three horizons. \n \n \n \n \n \n \n \n \n \n \n Contact: Silvi-Maria Gurova | IICT-BAS | silvimaria.gurova@iict.bas.bgTodor Gurov | IICT-BAS | todor.gurov@iict.bas.bgAleksandar Kirilov | IICT-BAS | alex.kirilov@iict.bas.bg URL:https://eurocc-bulgaria.bg/?tribe_events=from-building-to-city-energy-simulation-november-26-28-online ATTACH;FMTTYPE=image/png:https://eurocc-bulgaria.bg/wp-content/uploads/2026/03/From20to20Simulation_landscape_new.png END:VEVENT BEGIN:VEVENT DTSTART;VALUE=DATE:20251120 DTEND;VALUE=DATE:20251121 DTSTAMP:20261023T101421 CREATED:20251102T090053Z LAST-MODIFIED:20251112T133737Z UID:6628-1763596800-1763683199@eurocc-bulgaria.bg SUMMARY:Conference HPC 2025\, 20.11.2025 DESCRIPTION:NCC BULGARIA\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n NCC presenting the success story: NCC Bulgaria is the National Competence Centre for HPC and HPDA\, operating within the EuroCC3 network and EuroHPC ecosystem. Coordinated by the Institute of Information and Communication Technologies at the Bulgarian Academy of Sciences (IICT-BAS) in partnership with the Sofia University “St. Kliment Ohridski” (SU) and the University of National and World Economy (UNWE)\, it supports industry\, academia\, and public organizations through specialized training\, consulting\, technical expertise\, and access to national and European HPC resources. Scientific partners involved The Institute of Information and Communication Technologies (IICT-BAS) conducts fundamental and applied scientific research in the fields of high-performance computing (HPC)\, machine learning\, big data processing\, cybersecurity\, optimization\, financial modeling and mathematical modeling. The developed methods and technologies find application in medicine\, ecology\, industry\, transport\, financial analysis\, process optimization\, and digital transformation. Technical/scientific Challenge Constructing reliable investment portfolios requires balancing expected returns\, market volatility\, and potential losses under adverse market conditions. For a diversified universe of 32 global financial assets\, the large number of possible allocations makes constrained portfolio optimization computationally demanding. An additional challenge is that different risk measures\, such as the Sharpe ratio and Conditional Value-at-Risk (CVaR)\, may favour substantially different investment strategies. Portfolio decisions are also sensitive to the historical estimation period and the numerical sampling method. The research therefore investigates how Monte Carlo (MC) and randomized Quasi-Monte Carlo (QMC) algorithms can efficiently identify feasible portfolios while assessing numerical accuracy\, computational cost\, and sensitivity to market conditions. \n \n \n \n \n \n \n \n \n \n \n Solution Researchers at the Institute of Information and Communication Technologies of the Bulgarian Academy of Sciences (IICT-BAS) developed a reproducible Python-based framework for stochastic portfolio optimization using the HEMUS supercomputer. The methodology compares classical Monte Carlo sampling with Mersenne Twister and randomized Quasi-Monte Carlo methods based on scrambled Sobol and randomized Halton sequences. Candidate portfolios are generated under full-investment and long-only constraints\, with a maximum allocation of 30% per asset. We investigate two optimization objectives: maximizing the Sharpe ratio to improve risk-adjusted returns and minimizing historical CVaR to limit losses in the worst 5% of monthly observations. The analysis considers three historical estimation windows of 3\, 5\, and 10 years and five candidate budgets ranging from 128 to 2048. We evaluate each configuration through 50 independent randomized repetitions and compare it with deterministic numerical benchmarks using approximation errors\, portfolio characteristics\, and execution times. This computational framework enables systematic analysis of how sampling techniques\, risk objectives and historical periods influence portfolio selection. Scientific impact Methodology and Computational Framework The research presents a reproducible Python-based framework for stochastic portfolio optimization\, implemented on the HEMUS supercomputer at IICT-BAS. The study examines 32 global financial assets across 3-\, 5- and 10-year historical windows\, comparing Monte Carlo (with pseudorandom generator Mersenne Twister) and randomized Quasi-Monte Carlo (with scrambled Sobol and Halton) methods under Sharpe ratio maximization and CVaR minimization. Experiment 1 – Accuracy and Convergence of Stochastic Portfolio Search: Five sampling budgets (128–2048 candidate portfolios) and 50 independent repetitions per configuration were evaluated against deterministic numerical benchmarks. Increasing the number of candidates reduced approximation errors\, while differences between sampling methods remained modest\, with classical Monte Carlo generally requiring less execution time.Experiment 2 – Financial Risk – Return Comparison: The 5-year Sharpe-oriented portfolios achieved a mean annualized return of 9.15%\, while the 10-year CVaR-oriented portfolios recorded the lowest mean historical monthly CVaR of 4.40% and annualized volatility of 6.59%. These results highlight the trade-off between maximizing risk-adjusted returns and minimizing downside risk.Experiment 3 – Historical Window and Sampling Sensitivity: The 3-\, 5- and 10-year estimation windows produced different portfolio allocations and risk–return characteristics. The optimization objective and historical period generally had a greater influence on portfolio selection than the choice of sampling generator.Scientific Contribution The study demonstrates the potential of HPC-supported stochastic optimization for reproducible financial risk analysis and numerical benchmarking. The HEMUS supercomputer enabled extensive computational experiments\, providing a foundation for scalable portfolio optimization and future financial decision-support systems. Benefits:  Improved optimization accuracy: Larger sampling budgets reduce approximation errors relative to deterministic benchmarks.Risk-aware portfolio selection: Compares return-oriented Sharpe optimization with downside-risk-focused CVaR.Reliable numerical benchmarking: Evaluates accuracy\, stability and computational efficiency across MC/RQMC methods.Better understanding of market sensitivity: Reveals how historical estimation windows influence investment outcomes.HPC-enabled research: Supports repeated simulations and scalable numerical experimentation.Future financial applications: Provides a foundation for computational portfolio analysis and investment decision-support systems.\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Success story # Highlights: Keywords: HPC; Monte Carlo; Quasi-Monte Carlo; Portfolio Optimization; Sharpe Ratio; CVaR; Financial Risk.Technology: HEMUS Supercomputer; Python; Mersenne Twister; Scrambled Sobol and Halton; Stochastic Optimization.Industry sectors: Finance/Insurance; IT/HPC Systems\, Services & Software Providers.Scientific achievement: Development of a reproducible HPC-based framework for comparing stochastic portfolio optimization methods.Key results: Larger sampling budgets improved numerical accuracy\, while optimization objectives and historical windows significantly influenced portfolio risk–return characteristics.HPC contribution: HEMUS enabled repeated simulations\, algorithm benchmarking and computational performance analysis.\n \n \n \n \n Comparison of the 3 time horizons through Sharpe-ratio optimization. Among the sampled candidates\, the 5-year Sharpe solution provides the highest expected return and the strongest reward-to-risk trade-off\, at the cost of the highest volatility across the three horizons \n \n \n \n \n Figure 1. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 2. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 3. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Comparison of the 3 time horizons through CVaR (5%). At N=100\, Mean-CVaR shifts the selected portfolios toward lower-volatility regions. The 5-year solution achieves the highest expected return while slightly reducing volatility relative to the 3-year case\, whereas the 10-year solution is the most conservative in volatility terms. \n \n \n \n \n Figure 4. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 5. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 6. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 7. Sharpe vs CVaR: Cross-Horizon Comparison. Sharpe is higher at all three horizons. CVaR is lower at all three horizons. \n \n \n \n \n \n \n \n \n \n \n Contact: Silvi-Maria Gurova | IICT-BAS | silvimaria.gurova@iict.bas.bgTodor Gurov | IICT-BAS | todor.gurov@iict.bas.bgAleksandar Kirilov | IICT-BAS | alex.kirilov@iict.bas.bg URL:https://eurocc-bulgaria.bg/?tribe_events=conference-hpc-2025-20-11-2025 END:VEVENT BEGIN:VEVENT DTSTART;TZID=Europe/Helsinki:20251106T140000 DTEND;TZID=Europe/Helsinki:20251106T150000 DTSTAMP:20261023T101421 CREATED:20251101T110007Z LAST-MODIFIED:20251110T120052Z UID:6584-1762437600-1762441200@eurocc-bulgaria.bg SUMMARY:Minimal Coloring for Parallelization of Nearest-neighbour Operations on Cartesian Grids\, 6.11.2025 DESCRIPTION:NCC BULGARIA\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n NCC presenting the success story: NCC Bulgaria is the National Competence Centre for HPC and HPDA\, operating within the EuroCC3 network and EuroHPC ecosystem. Coordinated by the Institute of Information and Communication Technologies at the Bulgarian Academy of Sciences (IICT-BAS) in partnership with the Sofia University “St. Kliment Ohridski” (SU) and the University of National and World Economy (UNWE)\, it supports industry\, academia\, and public organizations through specialized training\, consulting\, technical expertise\, and access to national and European HPC resources. Scientific partners involved The Institute of Information and Communication Technologies (IICT-BAS) conducts fundamental and applied scientific research in the fields of high-performance computing (HPC)\, machine learning\, big data processing\, cybersecurity\, optimization\, financial modeling and mathematical modeling. The developed methods and technologies find application in medicine\, ecology\, industry\, transport\, financial analysis\, process optimization\, and digital transformation. Technical/scientific Challenge Constructing reliable investment portfolios requires balancing expected returns\, market volatility\, and potential losses under adverse market conditions. For a diversified universe of 32 global financial assets\, the large number of possible allocations makes constrained portfolio optimization computationally demanding. An additional challenge is that different risk measures\, such as the Sharpe ratio and Conditional Value-at-Risk (CVaR)\, may favour substantially different investment strategies. Portfolio decisions are also sensitive to the historical estimation period and the numerical sampling method. The research therefore investigates how Monte Carlo (MC) and randomized Quasi-Monte Carlo (QMC) algorithms can efficiently identify feasible portfolios while assessing numerical accuracy\, computational cost\, and sensitivity to market conditions. \n \n \n \n \n \n \n \n \n \n \n Solution Researchers at the Institute of Information and Communication Technologies of the Bulgarian Academy of Sciences (IICT-BAS) developed a reproducible Python-based framework for stochastic portfolio optimization using the HEMUS supercomputer. The methodology compares classical Monte Carlo sampling with Mersenne Twister and randomized Quasi-Monte Carlo methods based on scrambled Sobol and randomized Halton sequences. Candidate portfolios are generated under full-investment and long-only constraints\, with a maximum allocation of 30% per asset. We investigate two optimization objectives: maximizing the Sharpe ratio to improve risk-adjusted returns and minimizing historical CVaR to limit losses in the worst 5% of monthly observations. The analysis considers three historical estimation windows of 3\, 5\, and 10 years and five candidate budgets ranging from 128 to 2048. We evaluate each configuration through 50 independent randomized repetitions and compare it with deterministic numerical benchmarks using approximation errors\, portfolio characteristics\, and execution times. This computational framework enables systematic analysis of how sampling techniques\, risk objectives and historical periods influence portfolio selection. Scientific impact Methodology and Computational Framework The research presents a reproducible Python-based framework for stochastic portfolio optimization\, implemented on the HEMUS supercomputer at IICT-BAS. The study examines 32 global financial assets across 3-\, 5- and 10-year historical windows\, comparing Monte Carlo (with pseudorandom generator Mersenne Twister) and randomized Quasi-Monte Carlo (with scrambled Sobol and Halton) methods under Sharpe ratio maximization and CVaR minimization. Experiment 1 – Accuracy and Convergence of Stochastic Portfolio Search: Five sampling budgets (128–2048 candidate portfolios) and 50 independent repetitions per configuration were evaluated against deterministic numerical benchmarks. Increasing the number of candidates reduced approximation errors\, while differences between sampling methods remained modest\, with classical Monte Carlo generally requiring less execution time.Experiment 2 – Financial Risk – Return Comparison: The 5-year Sharpe-oriented portfolios achieved a mean annualized return of 9.15%\, while the 10-year CVaR-oriented portfolios recorded the lowest mean historical monthly CVaR of 4.40% and annualized volatility of 6.59%. These results highlight the trade-off between maximizing risk-adjusted returns and minimizing downside risk.Experiment 3 – Historical Window and Sampling Sensitivity: The 3-\, 5- and 10-year estimation windows produced different portfolio allocations and risk–return characteristics. The optimization objective and historical period generally had a greater influence on portfolio selection than the choice of sampling generator.Scientific Contribution The study demonstrates the potential of HPC-supported stochastic optimization for reproducible financial risk analysis and numerical benchmarking. The HEMUS supercomputer enabled extensive computational experiments\, providing a foundation for scalable portfolio optimization and future financial decision-support systems. Benefits:  Improved optimization accuracy: Larger sampling budgets reduce approximation errors relative to deterministic benchmarks.Risk-aware portfolio selection: Compares return-oriented Sharpe optimization with downside-risk-focused CVaR.Reliable numerical benchmarking: Evaluates accuracy\, stability and computational efficiency across MC/RQMC methods.Better understanding of market sensitivity: Reveals how historical estimation windows influence investment outcomes.HPC-enabled research: Supports repeated simulations and scalable numerical experimentation.Future financial applications: Provides a foundation for computational portfolio analysis and investment decision-support systems.\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Success story # Highlights: Keywords: HPC; Monte Carlo; Quasi-Monte Carlo; Portfolio Optimization; Sharpe Ratio; CVaR; Financial Risk.Technology: HEMUS Supercomputer; Python; Mersenne Twister; Scrambled Sobol and Halton; Stochastic Optimization.Industry sectors: Finance/Insurance; IT/HPC Systems\, Services & Software Providers.Scientific achievement: Development of a reproducible HPC-based framework for comparing stochastic portfolio optimization methods.Key results: Larger sampling budgets improved numerical accuracy\, while optimization objectives and historical windows significantly influenced portfolio risk–return characteristics.HPC contribution: HEMUS enabled repeated simulations\, algorithm benchmarking and computational performance analysis.\n \n \n \n \n Comparison of the 3 time horizons through Sharpe-ratio optimization. Among the sampled candidates\, the 5-year Sharpe solution provides the highest expected return and the strongest reward-to-risk trade-off\, at the cost of the highest volatility across the three horizons \n \n \n \n \n Figure 1. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 2. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 3. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Comparison of the 3 time horizons through CVaR (5%). At N=100\, Mean-CVaR shifts the selected portfolios toward lower-volatility regions. The 5-year solution achieves the highest expected return while slightly reducing volatility relative to the 3-year case\, whereas the 10-year solution is the most conservative in volatility terms. \n \n \n \n \n Figure 4. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 5. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 6. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 7. Sharpe vs CVaR: Cross-Horizon Comparison. Sharpe is higher at all three horizons. CVaR is lower at all three horizons. \n \n \n \n \n \n \n \n \n \n \n Contact: Silvi-Maria Gurova | IICT-BAS | silvimaria.gurova@iict.bas.bgTodor Gurov | IICT-BAS | todor.gurov@iict.bas.bgAleksandar Kirilov | IICT-BAS | alex.kirilov@iict.bas.bg URL:https://eurocc-bulgaria.bg/?tribe_events=minimal-coloring-for-parallelization-of-nearest-neighbour-operations-on-cartesian-grids-6-11-2025 END:VEVENT BEGIN:VEVENT DTSTART;TZID=Europe/Helsinki:20251106T140000 DTEND;TZID=Europe/Helsinki:20251106T150000 DTSTAMP:20261023T101421 CREATED:20251023T090016Z LAST-MODIFIED:20251112T133531Z UID:6624-1762437600-1762441200@eurocc-bulgaria.bg SUMMARY:HPC seminar series: Minimal coloring for parallelization of nearest-neighbor operations on Cartesian grids\, IICT\, 6.11.2025 DESCRIPTION:NCC BULGARIA\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n NCC presenting the success story: NCC Bulgaria is the National Competence Centre for HPC and HPDA\, operating within the EuroCC3 network and EuroHPC ecosystem. Coordinated by the Institute of Information and Communication Technologies at the Bulgarian Academy of Sciences (IICT-BAS) in partnership with the Sofia University “St. Kliment Ohridski” (SU) and the University of National and World Economy (UNWE)\, it supports industry\, academia\, and public organizations through specialized training\, consulting\, technical expertise\, and access to national and European HPC resources. Scientific partners involved The Institute of Information and Communication Technologies (IICT-BAS) conducts fundamental and applied scientific research in the fields of high-performance computing (HPC)\, machine learning\, big data processing\, cybersecurity\, optimization\, financial modeling and mathematical modeling. The developed methods and technologies find application in medicine\, ecology\, industry\, transport\, financial analysis\, process optimization\, and digital transformation. Technical/scientific Challenge Constructing reliable investment portfolios requires balancing expected returns\, market volatility\, and potential losses under adverse market conditions. For a diversified universe of 32 global financial assets\, the large number of possible allocations makes constrained portfolio optimization computationally demanding. An additional challenge is that different risk measures\, such as the Sharpe ratio and Conditional Value-at-Risk (CVaR)\, may favour substantially different investment strategies. Portfolio decisions are also sensitive to the historical estimation period and the numerical sampling method. The research therefore investigates how Monte Carlo (MC) and randomized Quasi-Monte Carlo (QMC) algorithms can efficiently identify feasible portfolios while assessing numerical accuracy\, computational cost\, and sensitivity to market conditions. \n \n \n \n \n \n \n \n \n \n \n Solution Researchers at the Institute of Information and Communication Technologies of the Bulgarian Academy of Sciences (IICT-BAS) developed a reproducible Python-based framework for stochastic portfolio optimization using the HEMUS supercomputer. The methodology compares classical Monte Carlo sampling with Mersenne Twister and randomized Quasi-Monte Carlo methods based on scrambled Sobol and randomized Halton sequences. Candidate portfolios are generated under full-investment and long-only constraints\, with a maximum allocation of 30% per asset. We investigate two optimization objectives: maximizing the Sharpe ratio to improve risk-adjusted returns and minimizing historical CVaR to limit losses in the worst 5% of monthly observations. The analysis considers three historical estimation windows of 3\, 5\, and 10 years and five candidate budgets ranging from 128 to 2048. We evaluate each configuration through 50 independent randomized repetitions and compare it with deterministic numerical benchmarks using approximation errors\, portfolio characteristics\, and execution times. This computational framework enables systematic analysis of how sampling techniques\, risk objectives and historical periods influence portfolio selection. Scientific impact Methodology and Computational Framework The research presents a reproducible Python-based framework for stochastic portfolio optimization\, implemented on the HEMUS supercomputer at IICT-BAS. The study examines 32 global financial assets across 3-\, 5- and 10-year historical windows\, comparing Monte Carlo (with pseudorandom generator Mersenne Twister) and randomized Quasi-Monte Carlo (with scrambled Sobol and Halton) methods under Sharpe ratio maximization and CVaR minimization. Experiment 1 – Accuracy and Convergence of Stochastic Portfolio Search: Five sampling budgets (128–2048 candidate portfolios) and 50 independent repetitions per configuration were evaluated against deterministic numerical benchmarks. Increasing the number of candidates reduced approximation errors\, while differences between sampling methods remained modest\, with classical Monte Carlo generally requiring less execution time.Experiment 2 – Financial Risk – Return Comparison: The 5-year Sharpe-oriented portfolios achieved a mean annualized return of 9.15%\, while the 10-year CVaR-oriented portfolios recorded the lowest mean historical monthly CVaR of 4.40% and annualized volatility of 6.59%. These results highlight the trade-off between maximizing risk-adjusted returns and minimizing downside risk.Experiment 3 – Historical Window and Sampling Sensitivity: The 3-\, 5- and 10-year estimation windows produced different portfolio allocations and risk–return characteristics. The optimization objective and historical period generally had a greater influence on portfolio selection than the choice of sampling generator.Scientific Contribution The study demonstrates the potential of HPC-supported stochastic optimization for reproducible financial risk analysis and numerical benchmarking. The HEMUS supercomputer enabled extensive computational experiments\, providing a foundation for scalable portfolio optimization and future financial decision-support systems. Benefits:  Improved optimization accuracy: Larger sampling budgets reduce approximation errors relative to deterministic benchmarks.Risk-aware portfolio selection: Compares return-oriented Sharpe optimization with downside-risk-focused CVaR.Reliable numerical benchmarking: Evaluates accuracy\, stability and computational efficiency across MC/RQMC methods.Better understanding of market sensitivity: Reveals how historical estimation windows influence investment outcomes.HPC-enabled research: Supports repeated simulations and scalable numerical experimentation.Future financial applications: Provides a foundation for computational portfolio analysis and investment decision-support systems.\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Success story # Highlights: Keywords: HPC; Monte Carlo; Quasi-Monte Carlo; Portfolio Optimization; Sharpe Ratio; CVaR; Financial Risk.Technology: HEMUS Supercomputer; Python; Mersenne Twister; Scrambled Sobol and Halton; Stochastic Optimization.Industry sectors: Finance/Insurance; IT/HPC Systems\, Services & Software Providers.Scientific achievement: Development of a reproducible HPC-based framework for comparing stochastic portfolio optimization methods.Key results: Larger sampling budgets improved numerical accuracy\, while optimization objectives and historical windows significantly influenced portfolio risk–return characteristics.HPC contribution: HEMUS enabled repeated simulations\, algorithm benchmarking and computational performance analysis.\n \n \n \n \n Comparison of the 3 time horizons through Sharpe-ratio optimization. Among the sampled candidates\, the 5-year Sharpe solution provides the highest expected return and the strongest reward-to-risk trade-off\, at the cost of the highest volatility across the three horizons \n \n \n \n \n Figure 1. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 2. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 3. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Comparison of the 3 time horizons through CVaR (5%). At N=100\, Mean-CVaR shifts the selected portfolios toward lower-volatility regions. The 5-year solution achieves the highest expected return while slightly reducing volatility relative to the 3-year case\, whereas the 10-year solution is the most conservative in volatility terms. \n \n \n \n \n Figure 4. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 5. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 6. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 7. Sharpe vs CVaR: Cross-Horizon Comparison. Sharpe is higher at all three horizons. CVaR is lower at all three horizons. \n \n \n \n \n \n \n \n \n \n \n Contact: Silvi-Maria Gurova | IICT-BAS | silvimaria.gurova@iict.bas.bgTodor Gurov | IICT-BAS | todor.gurov@iict.bas.bgAleksandar Kirilov | IICT-BAS | alex.kirilov@iict.bas.bg URL:https://eurocc-bulgaria.bg/?tribe_events=hpc-seminar-series-minimal-coloring-for-parallelization-of-nearest-neighbor-operations-on-cartesian-grids-iict-6-11-2025 END:VEVENT BEGIN:VEVENT DTSTART;VALUE=DATE:20251104 DTEND;VALUE=DATE:20251105 DTSTAMP:20261023T101421 CREATED:20251022T090031Z LAST-MODIFIED:20251112T133320Z UID:6620-1762214400-1762300799@eurocc-bulgaria.bg SUMMARY:Regional meeting “From science to innovations”\, Sofia Tech Park\, hall John Atanassov\, 4.11.2025 DESCRIPTION:NCC BULGARIA\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n NCC presenting the success story: NCC Bulgaria is the National Competence Centre for HPC and HPDA\, operating within the EuroCC3 network and EuroHPC ecosystem. Coordinated by the Institute of Information and Communication Technologies at the Bulgarian Academy of Sciences (IICT-BAS) in partnership with the Sofia University “St. Kliment Ohridski” (SU) and the University of National and World Economy (UNWE)\, it supports industry\, academia\, and public organizations through specialized training\, consulting\, technical expertise\, and access to national and European HPC resources. Scientific partners involved The Institute of Information and Communication Technologies (IICT-BAS) conducts fundamental and applied scientific research in the fields of high-performance computing (HPC)\, machine learning\, big data processing\, cybersecurity\, optimization\, financial modeling and mathematical modeling. The developed methods and technologies find application in medicine\, ecology\, industry\, transport\, financial analysis\, process optimization\, and digital transformation. Technical/scientific Challenge Constructing reliable investment portfolios requires balancing expected returns\, market volatility\, and potential losses under adverse market conditions. For a diversified universe of 32 global financial assets\, the large number of possible allocations makes constrained portfolio optimization computationally demanding. An additional challenge is that different risk measures\, such as the Sharpe ratio and Conditional Value-at-Risk (CVaR)\, may favour substantially different investment strategies. Portfolio decisions are also sensitive to the historical estimation period and the numerical sampling method. The research therefore investigates how Monte Carlo (MC) and randomized Quasi-Monte Carlo (QMC) algorithms can efficiently identify feasible portfolios while assessing numerical accuracy\, computational cost\, and sensitivity to market conditions. \n \n \n \n \n \n \n \n \n \n \n Solution Researchers at the Institute of Information and Communication Technologies of the Bulgarian Academy of Sciences (IICT-BAS) developed a reproducible Python-based framework for stochastic portfolio optimization using the HEMUS supercomputer. The methodology compares classical Monte Carlo sampling with Mersenne Twister and randomized Quasi-Monte Carlo methods based on scrambled Sobol and randomized Halton sequences. Candidate portfolios are generated under full-investment and long-only constraints\, with a maximum allocation of 30% per asset. We investigate two optimization objectives: maximizing the Sharpe ratio to improve risk-adjusted returns and minimizing historical CVaR to limit losses in the worst 5% of monthly observations. The analysis considers three historical estimation windows of 3\, 5\, and 10 years and five candidate budgets ranging from 128 to 2048. We evaluate each configuration through 50 independent randomized repetitions and compare it with deterministic numerical benchmarks using approximation errors\, portfolio characteristics\, and execution times. This computational framework enables systematic analysis of how sampling techniques\, risk objectives and historical periods influence portfolio selection. Scientific impact Methodology and Computational Framework The research presents a reproducible Python-based framework for stochastic portfolio optimization\, implemented on the HEMUS supercomputer at IICT-BAS. The study examines 32 global financial assets across 3-\, 5- and 10-year historical windows\, comparing Monte Carlo (with pseudorandom generator Mersenne Twister) and randomized Quasi-Monte Carlo (with scrambled Sobol and Halton) methods under Sharpe ratio maximization and CVaR minimization. Experiment 1 – Accuracy and Convergence of Stochastic Portfolio Search: Five sampling budgets (128–2048 candidate portfolios) and 50 independent repetitions per configuration were evaluated against deterministic numerical benchmarks. Increasing the number of candidates reduced approximation errors\, while differences between sampling methods remained modest\, with classical Monte Carlo generally requiring less execution time.Experiment 2 – Financial Risk – Return Comparison: The 5-year Sharpe-oriented portfolios achieved a mean annualized return of 9.15%\, while the 10-year CVaR-oriented portfolios recorded the lowest mean historical monthly CVaR of 4.40% and annualized volatility of 6.59%. These results highlight the trade-off between maximizing risk-adjusted returns and minimizing downside risk.Experiment 3 – Historical Window and Sampling Sensitivity: The 3-\, 5- and 10-year estimation windows produced different portfolio allocations and risk–return characteristics. The optimization objective and historical period generally had a greater influence on portfolio selection than the choice of sampling generator.Scientific Contribution The study demonstrates the potential of HPC-supported stochastic optimization for reproducible financial risk analysis and numerical benchmarking. The HEMUS supercomputer enabled extensive computational experiments\, providing a foundation for scalable portfolio optimization and future financial decision-support systems. Benefits:  Improved optimization accuracy: Larger sampling budgets reduce approximation errors relative to deterministic benchmarks.Risk-aware portfolio selection: Compares return-oriented Sharpe optimization with downside-risk-focused CVaR.Reliable numerical benchmarking: Evaluates accuracy\, stability and computational efficiency across MC/RQMC methods.Better understanding of market sensitivity: Reveals how historical estimation windows influence investment outcomes.HPC-enabled research: Supports repeated simulations and scalable numerical experimentation.Future financial applications: Provides a foundation for computational portfolio analysis and investment decision-support systems.\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Success story # Highlights: Keywords: HPC; Monte Carlo; Quasi-Monte Carlo; Portfolio Optimization; Sharpe Ratio; CVaR; Financial Risk.Technology: HEMUS Supercomputer; Python; Mersenne Twister; Scrambled Sobol and Halton; Stochastic Optimization.Industry sectors: Finance/Insurance; IT/HPC Systems\, Services & Software Providers.Scientific achievement: Development of a reproducible HPC-based framework for comparing stochastic portfolio optimization methods.Key results: Larger sampling budgets improved numerical accuracy\, while optimization objectives and historical windows significantly influenced portfolio risk–return characteristics.HPC contribution: HEMUS enabled repeated simulations\, algorithm benchmarking and computational performance analysis.\n \n \n \n \n Comparison of the 3 time horizons through Sharpe-ratio optimization. Among the sampled candidates\, the 5-year Sharpe solution provides the highest expected return and the strongest reward-to-risk trade-off\, at the cost of the highest volatility across the three horizons \n \n \n \n \n Figure 1. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 2. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 3. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Comparison of the 3 time horizons through CVaR (5%). At N=100\, Mean-CVaR shifts the selected portfolios toward lower-volatility regions. The 5-year solution achieves the highest expected return while slightly reducing volatility relative to the 3-year case\, whereas the 10-year solution is the most conservative in volatility terms. \n \n \n \n \n Figure 4. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 5. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 6. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 7. Sharpe vs CVaR: Cross-Horizon Comparison. Sharpe is higher at all three horizons. CVaR is lower at all three horizons. \n \n \n \n \n \n \n \n \n \n \n Contact: Silvi-Maria Gurova | IICT-BAS | silvimaria.gurova@iict.bas.bgTodor Gurov | IICT-BAS | todor.gurov@iict.bas.bgAleksandar Kirilov | IICT-BAS | alex.kirilov@iict.bas.bg URL:https://eurocc-bulgaria.bg/?tribe_events=regional-meeting-from-science-to-innovations-sofia-tech-park-hall-john-atanassov-4-11-2025 END:VEVENT BEGIN:VEVENT DTSTART;TZID=Europe/Helsinki:20251028T100000 DTEND;TZID=Europe/Helsinki:20251028T130000 DTSTAMP:20261023T101421 CREATED:20251012T110024Z LAST-MODIFIED:20260423T112252Z UID:6616-1761645600-1761656400@eurocc-bulgaria.bg SUMMARY:Training ”Supercomputer HEMUS: Features and Development Tools”\, IICT\, 28.10.2025 DESCRIPTION:NCC BULGARIA\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n NCC presenting the success story: NCC Bulgaria is the National Competence Centre for HPC and HPDA\, operating within the EuroCC3 network and EuroHPC ecosystem. Coordinated by the Institute of Information and Communication Technologies at the Bulgarian Academy of Sciences (IICT-BAS) in partnership with the Sofia University “St. Kliment Ohridski” (SU) and the University of National and World Economy (UNWE)\, it supports industry\, academia\, and public organizations through specialized training\, consulting\, technical expertise\, and access to national and European HPC resources. Scientific partners involved The Institute of Information and Communication Technologies (IICT-BAS) conducts fundamental and applied scientific research in the fields of high-performance computing (HPC)\, machine learning\, big data processing\, cybersecurity\, optimization\, financial modeling and mathematical modeling. The developed methods and technologies find application in medicine\, ecology\, industry\, transport\, financial analysis\, process optimization\, and digital transformation. Technical/scientific Challenge Constructing reliable investment portfolios requires balancing expected returns\, market volatility\, and potential losses under adverse market conditions. For a diversified universe of 32 global financial assets\, the large number of possible allocations makes constrained portfolio optimization computationally demanding. An additional challenge is that different risk measures\, such as the Sharpe ratio and Conditional Value-at-Risk (CVaR)\, may favour substantially different investment strategies. Portfolio decisions are also sensitive to the historical estimation period and the numerical sampling method. The research therefore investigates how Monte Carlo (MC) and randomized Quasi-Monte Carlo (QMC) algorithms can efficiently identify feasible portfolios while assessing numerical accuracy\, computational cost\, and sensitivity to market conditions. \n \n \n \n \n \n \n \n \n \n \n Solution Researchers at the Institute of Information and Communication Technologies of the Bulgarian Academy of Sciences (IICT-BAS) developed a reproducible Python-based framework for stochastic portfolio optimization using the HEMUS supercomputer. The methodology compares classical Monte Carlo sampling with Mersenne Twister and randomized Quasi-Monte Carlo methods based on scrambled Sobol and randomized Halton sequences. Candidate portfolios are generated under full-investment and long-only constraints\, with a maximum allocation of 30% per asset. We investigate two optimization objectives: maximizing the Sharpe ratio to improve risk-adjusted returns and minimizing historical CVaR to limit losses in the worst 5% of monthly observations. The analysis considers three historical estimation windows of 3\, 5\, and 10 years and five candidate budgets ranging from 128 to 2048. We evaluate each configuration through 50 independent randomized repetitions and compare it with deterministic numerical benchmarks using approximation errors\, portfolio characteristics\, and execution times. This computational framework enables systematic analysis of how sampling techniques\, risk objectives and historical periods influence portfolio selection. Scientific impact Methodology and Computational Framework The research presents a reproducible Python-based framework for stochastic portfolio optimization\, implemented on the HEMUS supercomputer at IICT-BAS. The study examines 32 global financial assets across 3-\, 5- and 10-year historical windows\, comparing Monte Carlo (with pseudorandom generator Mersenne Twister) and randomized Quasi-Monte Carlo (with scrambled Sobol and Halton) methods under Sharpe ratio maximization and CVaR minimization. Experiment 1 – Accuracy and Convergence of Stochastic Portfolio Search: Five sampling budgets (128–2048 candidate portfolios) and 50 independent repetitions per configuration were evaluated against deterministic numerical benchmarks. Increasing the number of candidates reduced approximation errors\, while differences between sampling methods remained modest\, with classical Monte Carlo generally requiring less execution time.Experiment 2 – Financial Risk – Return Comparison: The 5-year Sharpe-oriented portfolios achieved a mean annualized return of 9.15%\, while the 10-year CVaR-oriented portfolios recorded the lowest mean historical monthly CVaR of 4.40% and annualized volatility of 6.59%. These results highlight the trade-off between maximizing risk-adjusted returns and minimizing downside risk.Experiment 3 – Historical Window and Sampling Sensitivity: The 3-\, 5- and 10-year estimation windows produced different portfolio allocations and risk–return characteristics. The optimization objective and historical period generally had a greater influence on portfolio selection than the choice of sampling generator.Scientific Contribution The study demonstrates the potential of HPC-supported stochastic optimization for reproducible financial risk analysis and numerical benchmarking. The HEMUS supercomputer enabled extensive computational experiments\, providing a foundation for scalable portfolio optimization and future financial decision-support systems. Benefits:  Improved optimization accuracy: Larger sampling budgets reduce approximation errors relative to deterministic benchmarks.Risk-aware portfolio selection: Compares return-oriented Sharpe optimization with downside-risk-focused CVaR.Reliable numerical benchmarking: Evaluates accuracy\, stability and computational efficiency across MC/RQMC methods.Better understanding of market sensitivity: Reveals how historical estimation windows influence investment outcomes.HPC-enabled research: Supports repeated simulations and scalable numerical experimentation.Future financial applications: Provides a foundation for computational portfolio analysis and investment decision-support systems.\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Success story # Highlights: Keywords: HPC; Monte Carlo; Quasi-Monte Carlo; Portfolio Optimization; Sharpe Ratio; CVaR; Financial Risk.Technology: HEMUS Supercomputer; Python; Mersenne Twister; Scrambled Sobol and Halton; Stochastic Optimization.Industry sectors: Finance/Insurance; IT/HPC Systems\, Services & Software Providers.Scientific achievement: Development of a reproducible HPC-based framework for comparing stochastic portfolio optimization methods.Key results: Larger sampling budgets improved numerical accuracy\, while optimization objectives and historical windows significantly influenced portfolio risk–return characteristics.HPC contribution: HEMUS enabled repeated simulations\, algorithm benchmarking and computational performance analysis.\n \n \n \n \n Comparison of the 3 time horizons through Sharpe-ratio optimization. Among the sampled candidates\, the 5-year Sharpe solution provides the highest expected return and the strongest reward-to-risk trade-off\, at the cost of the highest volatility across the three horizons \n \n \n \n \n Figure 1. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 2. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 3. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Comparison of the 3 time horizons through CVaR (5%). At N=100\, Mean-CVaR shifts the selected portfolios toward lower-volatility regions. The 5-year solution achieves the highest expected return while slightly reducing volatility relative to the 3-year case\, whereas the 10-year solution is the most conservative in volatility terms. \n \n \n \n \n Figure 4. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 5. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 6. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 7. Sharpe vs CVaR: Cross-Horizon Comparison. Sharpe is higher at all three horizons. CVaR is lower at all three horizons. \n \n \n \n \n \n \n \n \n \n \n Contact: Silvi-Maria Gurova | IICT-BAS | silvimaria.gurova@iict.bas.bgTodor Gurov | IICT-BAS | todor.gurov@iict.bas.bgAleksandar Kirilov | IICT-BAS | alex.kirilov@iict.bas.bg URL:https://eurocc-bulgaria.bg/?tribe_events=training-supercomputer-hemus-features-and-development-tools-iict-28-10-2025 END:VEVENT BEGIN:VEVENT DTSTART;TZID=Europe/Helsinki:20251028T090000 DTEND;TZID=Europe/Helsinki:20251028T130000 DTSTAMP:20261023T101421 CREATED:20251012T090007Z LAST-MODIFIED:20251112T132800Z UID:6612-1761642000-1761656400@eurocc-bulgaria.bg SUMMARY:Open Doors Day at IICT: “Digital transformation and AI”\, 28.10.2025 DESCRIPTION:NCC BULGARIA\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n NCC presenting the success story: NCC Bulgaria is the National Competence Centre for HPC and HPDA\, operating within the EuroCC3 network and EuroHPC ecosystem. Coordinated by the Institute of Information and Communication Technologies at the Bulgarian Academy of Sciences (IICT-BAS) in partnership with the Sofia University “St. Kliment Ohridski” (SU) and the University of National and World Economy (UNWE)\, it supports industry\, academia\, and public organizations through specialized training\, consulting\, technical expertise\, and access to national and European HPC resources. Scientific partners involved The Institute of Information and Communication Technologies (IICT-BAS) conducts fundamental and applied scientific research in the fields of high-performance computing (HPC)\, machine learning\, big data processing\, cybersecurity\, optimization\, financial modeling and mathematical modeling. The developed methods and technologies find application in medicine\, ecology\, industry\, transport\, financial analysis\, process optimization\, and digital transformation. Technical/scientific Challenge Constructing reliable investment portfolios requires balancing expected returns\, market volatility\, and potential losses under adverse market conditions. For a diversified universe of 32 global financial assets\, the large number of possible allocations makes constrained portfolio optimization computationally demanding. An additional challenge is that different risk measures\, such as the Sharpe ratio and Conditional Value-at-Risk (CVaR)\, may favour substantially different investment strategies. Portfolio decisions are also sensitive to the historical estimation period and the numerical sampling method. The research therefore investigates how Monte Carlo (MC) and randomized Quasi-Monte Carlo (QMC) algorithms can efficiently identify feasible portfolios while assessing numerical accuracy\, computational cost\, and sensitivity to market conditions. \n \n \n \n \n \n \n \n \n \n \n Solution Researchers at the Institute of Information and Communication Technologies of the Bulgarian Academy of Sciences (IICT-BAS) developed a reproducible Python-based framework for stochastic portfolio optimization using the HEMUS supercomputer. The methodology compares classical Monte Carlo sampling with Mersenne Twister and randomized Quasi-Monte Carlo methods based on scrambled Sobol and randomized Halton sequences. Candidate portfolios are generated under full-investment and long-only constraints\, with a maximum allocation of 30% per asset. We investigate two optimization objectives: maximizing the Sharpe ratio to improve risk-adjusted returns and minimizing historical CVaR to limit losses in the worst 5% of monthly observations. The analysis considers three historical estimation windows of 3\, 5\, and 10 years and five candidate budgets ranging from 128 to 2048. We evaluate each configuration through 50 independent randomized repetitions and compare it with deterministic numerical benchmarks using approximation errors\, portfolio characteristics\, and execution times. This computational framework enables systematic analysis of how sampling techniques\, risk objectives and historical periods influence portfolio selection. Scientific impact Methodology and Computational Framework The research presents a reproducible Python-based framework for stochastic portfolio optimization\, implemented on the HEMUS supercomputer at IICT-BAS. The study examines 32 global financial assets across 3-\, 5- and 10-year historical windows\, comparing Monte Carlo (with pseudorandom generator Mersenne Twister) and randomized Quasi-Monte Carlo (with scrambled Sobol and Halton) methods under Sharpe ratio maximization and CVaR minimization. Experiment 1 – Accuracy and Convergence of Stochastic Portfolio Search: Five sampling budgets (128–2048 candidate portfolios) and 50 independent repetitions per configuration were evaluated against deterministic numerical benchmarks. Increasing the number of candidates reduced approximation errors\, while differences between sampling methods remained modest\, with classical Monte Carlo generally requiring less execution time.Experiment 2 – Financial Risk – Return Comparison: The 5-year Sharpe-oriented portfolios achieved a mean annualized return of 9.15%\, while the 10-year CVaR-oriented portfolios recorded the lowest mean historical monthly CVaR of 4.40% and annualized volatility of 6.59%. These results highlight the trade-off between maximizing risk-adjusted returns and minimizing downside risk.Experiment 3 – Historical Window and Sampling Sensitivity: The 3-\, 5- and 10-year estimation windows produced different portfolio allocations and risk–return characteristics. The optimization objective and historical period generally had a greater influence on portfolio selection than the choice of sampling generator.Scientific Contribution The study demonstrates the potential of HPC-supported stochastic optimization for reproducible financial risk analysis and numerical benchmarking. The HEMUS supercomputer enabled extensive computational experiments\, providing a foundation for scalable portfolio optimization and future financial decision-support systems. Benefits:  Improved optimization accuracy: Larger sampling budgets reduce approximation errors relative to deterministic benchmarks.Risk-aware portfolio selection: Compares return-oriented Sharpe optimization with downside-risk-focused CVaR.Reliable numerical benchmarking: Evaluates accuracy\, stability and computational efficiency across MC/RQMC methods.Better understanding of market sensitivity: Reveals how historical estimation windows influence investment outcomes.HPC-enabled research: Supports repeated simulations and scalable numerical experimentation.Future financial applications: Provides a foundation for computational portfolio analysis and investment decision-support systems.\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Success story # Highlights: Keywords: HPC; Monte Carlo; Quasi-Monte Carlo; Portfolio Optimization; Sharpe Ratio; CVaR; Financial Risk.Technology: HEMUS Supercomputer; Python; Mersenne Twister; Scrambled Sobol and Halton; Stochastic Optimization.Industry sectors: Finance/Insurance; IT/HPC Systems\, Services & Software Providers.Scientific achievement: Development of a reproducible HPC-based framework for comparing stochastic portfolio optimization methods.Key results: Larger sampling budgets improved numerical accuracy\, while optimization objectives and historical windows significantly influenced portfolio risk–return characteristics.HPC contribution: HEMUS enabled repeated simulations\, algorithm benchmarking and computational performance analysis.\n \n \n \n \n Comparison of the 3 time horizons through Sharpe-ratio optimization. Among the sampled candidates\, the 5-year Sharpe solution provides the highest expected return and the strongest reward-to-risk trade-off\, at the cost of the highest volatility across the three horizons \n \n \n \n \n Figure 1. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 2. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 3. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Comparison of the 3 time horizons through CVaR (5%). At N=100\, Mean-CVaR shifts the selected portfolios toward lower-volatility regions. The 5-year solution achieves the highest expected return while slightly reducing volatility relative to the 3-year case\, whereas the 10-year solution is the most conservative in volatility terms. \n \n \n \n \n Figure 4. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 5. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 6. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 7. Sharpe vs CVaR: Cross-Horizon Comparison. Sharpe is higher at all three horizons. CVaR is lower at all three horizons. \n \n \n \n \n \n \n \n \n \n \n Contact: Silvi-Maria Gurova | IICT-BAS | silvimaria.gurova@iict.bas.bgTodor Gurov | IICT-BAS | todor.gurov@iict.bas.bgAleksandar Kirilov | IICT-BAS | alex.kirilov@iict.bas.bg URL:https://eurocc-bulgaria.bg/?tribe_events=open-doors-day-at-iict-digital-transformation-and-ai-28-10-2025 END:VEVENT BEGIN:VEVENT DTSTART;VALUE=DATE:20251022 DTEND;VALUE=DATE:20251023 DTSTAMP:20261023T101421 CREATED:20251010T090005Z LAST-MODIFIED:20251112T132555Z UID:6608-1761091200-1761177599@eurocc-bulgaria.bg SUMMARY:Emerging Trends in High-Performance Computing and their Cybersecurity Implications\, International seminar with Round Table\, Central Military Club\, Sofia\, Bulgaria\, 22.10.2025 DESCRIPTION:NCC BULGARIA\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n NCC presenting the success story: NCC Bulgaria is the National Competence Centre for HPC and HPDA\, operating within the EuroCC3 network and EuroHPC ecosystem. Coordinated by the Institute of Information and Communication Technologies at the Bulgarian Academy of Sciences (IICT-BAS) in partnership with the Sofia University “St. Kliment Ohridski” (SU) and the University of National and World Economy (UNWE)\, it supports industry\, academia\, and public organizations through specialized training\, consulting\, technical expertise\, and access to national and European HPC resources. Scientific partners involved The Institute of Information and Communication Technologies (IICT-BAS) conducts fundamental and applied scientific research in the fields of high-performance computing (HPC)\, machine learning\, big data processing\, cybersecurity\, optimization\, financial modeling and mathematical modeling. The developed methods and technologies find application in medicine\, ecology\, industry\, transport\, financial analysis\, process optimization\, and digital transformation. Technical/scientific Challenge Constructing reliable investment portfolios requires balancing expected returns\, market volatility\, and potential losses under adverse market conditions. For a diversified universe of 32 global financial assets\, the large number of possible allocations makes constrained portfolio optimization computationally demanding. An additional challenge is that different risk measures\, such as the Sharpe ratio and Conditional Value-at-Risk (CVaR)\, may favour substantially different investment strategies. Portfolio decisions are also sensitive to the historical estimation period and the numerical sampling method. The research therefore investigates how Monte Carlo (MC) and randomized Quasi-Monte Carlo (QMC) algorithms can efficiently identify feasible portfolios while assessing numerical accuracy\, computational cost\, and sensitivity to market conditions. \n \n \n \n \n \n \n \n \n \n \n Solution Researchers at the Institute of Information and Communication Technologies of the Bulgarian Academy of Sciences (IICT-BAS) developed a reproducible Python-based framework for stochastic portfolio optimization using the HEMUS supercomputer. The methodology compares classical Monte Carlo sampling with Mersenne Twister and randomized Quasi-Monte Carlo methods based on scrambled Sobol and randomized Halton sequences. Candidate portfolios are generated under full-investment and long-only constraints\, with a maximum allocation of 30% per asset. We investigate two optimization objectives: maximizing the Sharpe ratio to improve risk-adjusted returns and minimizing historical CVaR to limit losses in the worst 5% of monthly observations. The analysis considers three historical estimation windows of 3\, 5\, and 10 years and five candidate budgets ranging from 128 to 2048. We evaluate each configuration through 50 independent randomized repetitions and compare it with deterministic numerical benchmarks using approximation errors\, portfolio characteristics\, and execution times. This computational framework enables systematic analysis of how sampling techniques\, risk objectives and historical periods influence portfolio selection. Scientific impact Methodology and Computational Framework The research presents a reproducible Python-based framework for stochastic portfolio optimization\, implemented on the HEMUS supercomputer at IICT-BAS. The study examines 32 global financial assets across 3-\, 5- and 10-year historical windows\, comparing Monte Carlo (with pseudorandom generator Mersenne Twister) and randomized Quasi-Monte Carlo (with scrambled Sobol and Halton) methods under Sharpe ratio maximization and CVaR minimization. Experiment 1 – Accuracy and Convergence of Stochastic Portfolio Search: Five sampling budgets (128–2048 candidate portfolios) and 50 independent repetitions per configuration were evaluated against deterministic numerical benchmarks. Increasing the number of candidates reduced approximation errors\, while differences between sampling methods remained modest\, with classical Monte Carlo generally requiring less execution time.Experiment 2 – Financial Risk – Return Comparison: The 5-year Sharpe-oriented portfolios achieved a mean annualized return of 9.15%\, while the 10-year CVaR-oriented portfolios recorded the lowest mean historical monthly CVaR of 4.40% and annualized volatility of 6.59%. These results highlight the trade-off between maximizing risk-adjusted returns and minimizing downside risk.Experiment 3 – Historical Window and Sampling Sensitivity: The 3-\, 5- and 10-year estimation windows produced different portfolio allocations and risk–return characteristics. The optimization objective and historical period generally had a greater influence on portfolio selection than the choice of sampling generator.Scientific Contribution The study demonstrates the potential of HPC-supported stochastic optimization for reproducible financial risk analysis and numerical benchmarking. The HEMUS supercomputer enabled extensive computational experiments\, providing a foundation for scalable portfolio optimization and future financial decision-support systems. Benefits:  Improved optimization accuracy: Larger sampling budgets reduce approximation errors relative to deterministic benchmarks.Risk-aware portfolio selection: Compares return-oriented Sharpe optimization with downside-risk-focused CVaR.Reliable numerical benchmarking: Evaluates accuracy\, stability and computational efficiency across MC/RQMC methods.Better understanding of market sensitivity: Reveals how historical estimation windows influence investment outcomes.HPC-enabled research: Supports repeated simulations and scalable numerical experimentation.Future financial applications: Provides a foundation for computational portfolio analysis and investment decision-support systems.\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Success story # Highlights: Keywords: HPC; Monte Carlo; Quasi-Monte Carlo; Portfolio Optimization; Sharpe Ratio; CVaR; Financial Risk.Technology: HEMUS Supercomputer; Python; Mersenne Twister; Scrambled Sobol and Halton; Stochastic Optimization.Industry sectors: Finance/Insurance; IT/HPC Systems\, Services & Software Providers.Scientific achievement: Development of a reproducible HPC-based framework for comparing stochastic portfolio optimization methods.Key results: Larger sampling budgets improved numerical accuracy\, while optimization objectives and historical windows significantly influenced portfolio risk–return characteristics.HPC contribution: HEMUS enabled repeated simulations\, algorithm benchmarking and computational performance analysis.\n \n \n \n \n Comparison of the 3 time horizons through Sharpe-ratio optimization. Among the sampled candidates\, the 5-year Sharpe solution provides the highest expected return and the strongest reward-to-risk trade-off\, at the cost of the highest volatility across the three horizons \n \n \n \n \n Figure 1. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 2. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 3. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Comparison of the 3 time horizons through CVaR (5%). At N=100\, Mean-CVaR shifts the selected portfolios toward lower-volatility regions. The 5-year solution achieves the highest expected return while slightly reducing volatility relative to the 3-year case\, whereas the 10-year solution is the most conservative in volatility terms. \n \n \n \n \n Figure 4. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 5. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 6. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 7. Sharpe vs CVaR: Cross-Horizon Comparison. Sharpe is higher at all three horizons. CVaR is lower at all three horizons. \n \n \n \n \n \n \n \n \n \n \n Contact: Silvi-Maria Gurova | IICT-BAS | silvimaria.gurova@iict.bas.bgTodor Gurov | IICT-BAS | todor.gurov@iict.bas.bgAleksandar Kirilov | IICT-BAS | alex.kirilov@iict.bas.bg URL:https://eurocc-bulgaria.bg/?tribe_events=emerging-trends-in-high-performance-computing-and-their-cybersecurity-implications-international-seminar-with-round-table-central-military-club-sofia-bulgaria-22-10-2025 END:VEVENT BEGIN:VEVENT DTSTART;TZID=Europe/Helsinki:20251014T140000 DTEND;TZID=Europe/Helsinki:20251014T150000 DTSTAMP:20261023T101421 CREATED:20251008T090033Z LAST-MODIFIED:20251112T132359Z UID:6604-1760450400-1760454000@eurocc-bulgaria.bg SUMMARY:HPC seminar series: Quantum simulators\, Lecturer E. Atanassov (IICT) DESCRIPTION:NCC BULGARIA\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n NCC presenting the success story: NCC Bulgaria is the National Competence Centre for HPC and HPDA\, operating within the EuroCC3 network and EuroHPC ecosystem. Coordinated by the Institute of Information and Communication Technologies at the Bulgarian Academy of Sciences (IICT-BAS) in partnership with the Sofia University “St. Kliment Ohridski” (SU) and the University of National and World Economy (UNWE)\, it supports industry\, academia\, and public organizations through specialized training\, consulting\, technical expertise\, and access to national and European HPC resources. Scientific partners involved The Institute of Information and Communication Technologies (IICT-BAS) conducts fundamental and applied scientific research in the fields of high-performance computing (HPC)\, machine learning\, big data processing\, cybersecurity\, optimization\, financial modeling and mathematical modeling. The developed methods and technologies find application in medicine\, ecology\, industry\, transport\, financial analysis\, process optimization\, and digital transformation. Technical/scientific Challenge Constructing reliable investment portfolios requires balancing expected returns\, market volatility\, and potential losses under adverse market conditions. For a diversified universe of 32 global financial assets\, the large number of possible allocations makes constrained portfolio optimization computationally demanding. An additional challenge is that different risk measures\, such as the Sharpe ratio and Conditional Value-at-Risk (CVaR)\, may favour substantially different investment strategies. Portfolio decisions are also sensitive to the historical estimation period and the numerical sampling method. The research therefore investigates how Monte Carlo (MC) and randomized Quasi-Monte Carlo (QMC) algorithms can efficiently identify feasible portfolios while assessing numerical accuracy\, computational cost\, and sensitivity to market conditions. \n \n \n \n \n \n \n \n \n \n \n Solution Researchers at the Institute of Information and Communication Technologies of the Bulgarian Academy of Sciences (IICT-BAS) developed a reproducible Python-based framework for stochastic portfolio optimization using the HEMUS supercomputer. The methodology compares classical Monte Carlo sampling with Mersenne Twister and randomized Quasi-Monte Carlo methods based on scrambled Sobol and randomized Halton sequences. Candidate portfolios are generated under full-investment and long-only constraints\, with a maximum allocation of 30% per asset. We investigate two optimization objectives: maximizing the Sharpe ratio to improve risk-adjusted returns and minimizing historical CVaR to limit losses in the worst 5% of monthly observations. The analysis considers three historical estimation windows of 3\, 5\, and 10 years and five candidate budgets ranging from 128 to 2048. We evaluate each configuration through 50 independent randomized repetitions and compare it with deterministic numerical benchmarks using approximation errors\, portfolio characteristics\, and execution times. This computational framework enables systematic analysis of how sampling techniques\, risk objectives and historical periods influence portfolio selection. Scientific impact Methodology and Computational Framework The research presents a reproducible Python-based framework for stochastic portfolio optimization\, implemented on the HEMUS supercomputer at IICT-BAS. The study examines 32 global financial assets across 3-\, 5- and 10-year historical windows\, comparing Monte Carlo (with pseudorandom generator Mersenne Twister) and randomized Quasi-Monte Carlo (with scrambled Sobol and Halton) methods under Sharpe ratio maximization and CVaR minimization. Experiment 1 – Accuracy and Convergence of Stochastic Portfolio Search: Five sampling budgets (128–2048 candidate portfolios) and 50 independent repetitions per configuration were evaluated against deterministic numerical benchmarks. Increasing the number of candidates reduced approximation errors\, while differences between sampling methods remained modest\, with classical Monte Carlo generally requiring less execution time.Experiment 2 – Financial Risk – Return Comparison: The 5-year Sharpe-oriented portfolios achieved a mean annualized return of 9.15%\, while the 10-year CVaR-oriented portfolios recorded the lowest mean historical monthly CVaR of 4.40% and annualized volatility of 6.59%. These results highlight the trade-off between maximizing risk-adjusted returns and minimizing downside risk.Experiment 3 – Historical Window and Sampling Sensitivity: The 3-\, 5- and 10-year estimation windows produced different portfolio allocations and risk–return characteristics. The optimization objective and historical period generally had a greater influence on portfolio selection than the choice of sampling generator.Scientific Contribution The study demonstrates the potential of HPC-supported stochastic optimization for reproducible financial risk analysis and numerical benchmarking. The HEMUS supercomputer enabled extensive computational experiments\, providing a foundation for scalable portfolio optimization and future financial decision-support systems. Benefits:  Improved optimization accuracy: Larger sampling budgets reduce approximation errors relative to deterministic benchmarks.Risk-aware portfolio selection: Compares return-oriented Sharpe optimization with downside-risk-focused CVaR.Reliable numerical benchmarking: Evaluates accuracy\, stability and computational efficiency across MC/RQMC methods.Better understanding of market sensitivity: Reveals how historical estimation windows influence investment outcomes.HPC-enabled research: Supports repeated simulations and scalable numerical experimentation.Future financial applications: Provides a foundation for computational portfolio analysis and investment decision-support systems.\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Success story # Highlights: Keywords: HPC; Monte Carlo; Quasi-Monte Carlo; Portfolio Optimization; Sharpe Ratio; CVaR; Financial Risk.Technology: HEMUS Supercomputer; Python; Mersenne Twister; Scrambled Sobol and Halton; Stochastic Optimization.Industry sectors: Finance/Insurance; IT/HPC Systems\, Services & Software Providers.Scientific achievement: Development of a reproducible HPC-based framework for comparing stochastic portfolio optimization methods.Key results: Larger sampling budgets improved numerical accuracy\, while optimization objectives and historical windows significantly influenced portfolio risk–return characteristics.HPC contribution: HEMUS enabled repeated simulations\, algorithm benchmarking and computational performance analysis.\n \n \n \n \n Comparison of the 3 time horizons through Sharpe-ratio optimization. Among the sampled candidates\, the 5-year Sharpe solution provides the highest expected return and the strongest reward-to-risk trade-off\, at the cost of the highest volatility across the three horizons \n \n \n \n \n Figure 1. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 2. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 3. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Comparison of the 3 time horizons through CVaR (5%). At N=100\, Mean-CVaR shifts the selected portfolios toward lower-volatility regions. The 5-year solution achieves the highest expected return while slightly reducing volatility relative to the 3-year case\, whereas the 10-year solution is the most conservative in volatility terms. \n \n \n \n \n Figure 4. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 5. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 6. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 7. Sharpe vs CVaR: Cross-Horizon Comparison. Sharpe is higher at all three horizons. CVaR is lower at all three horizons. \n \n \n \n \n \n \n \n \n \n \n Contact: Silvi-Maria Gurova | IICT-BAS | silvimaria.gurova@iict.bas.bgTodor Gurov | IICT-BAS | todor.gurov@iict.bas.bgAleksandar Kirilov | IICT-BAS | alex.kirilov@iict.bas.bg URL:https://eurocc-bulgaria.bg/?tribe_events=hpc-seminar-series-quantum-simulators-lecturer-e-atanassov-iict END:VEVENT BEGIN:VEVENT DTSTART;VALUE=DATE:20251002 DTEND;VALUE=DATE:20251003 DTSTAMP:20261023T101421 CREATED:20250922T090008Z LAST-MODIFIED:20251112T132016Z UID:6599-1759363200-1759449599@eurocc-bulgaria.bg SUMMARY:Training on HPC Fundamentals with Cybersecurity Applications\, Plovdiv\, 2.10.2025 DESCRIPTION:NCC BULGARIA\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n NCC presenting the success story: NCC Bulgaria is the National Competence Centre for HPC and HPDA\, operating within the EuroCC3 network and EuroHPC ecosystem. Coordinated by the Institute of Information and Communication Technologies at the Bulgarian Academy of Sciences (IICT-BAS) in partnership with the Sofia University “St. Kliment Ohridski” (SU) and the University of National and World Economy (UNWE)\, it supports industry\, academia\, and public organizations through specialized training\, consulting\, technical expertise\, and access to national and European HPC resources. Scientific partners involved The Institute of Information and Communication Technologies (IICT-BAS) conducts fundamental and applied scientific research in the fields of high-performance computing (HPC)\, machine learning\, big data processing\, cybersecurity\, optimization\, financial modeling and mathematical modeling. The developed methods and technologies find application in medicine\, ecology\, industry\, transport\, financial analysis\, process optimization\, and digital transformation. Technical/scientific Challenge Constructing reliable investment portfolios requires balancing expected returns\, market volatility\, and potential losses under adverse market conditions. For a diversified universe of 32 global financial assets\, the large number of possible allocations makes constrained portfolio optimization computationally demanding. An additional challenge is that different risk measures\, such as the Sharpe ratio and Conditional Value-at-Risk (CVaR)\, may favour substantially different investment strategies. Portfolio decisions are also sensitive to the historical estimation period and the numerical sampling method. The research therefore investigates how Monte Carlo (MC) and randomized Quasi-Monte Carlo (QMC) algorithms can efficiently identify feasible portfolios while assessing numerical accuracy\, computational cost\, and sensitivity to market conditions. \n \n \n \n \n \n \n \n \n \n \n Solution Researchers at the Institute of Information and Communication Technologies of the Bulgarian Academy of Sciences (IICT-BAS) developed a reproducible Python-based framework for stochastic portfolio optimization using the HEMUS supercomputer. The methodology compares classical Monte Carlo sampling with Mersenne Twister and randomized Quasi-Monte Carlo methods based on scrambled Sobol and randomized Halton sequences. Candidate portfolios are generated under full-investment and long-only constraints\, with a maximum allocation of 30% per asset. We investigate two optimization objectives: maximizing the Sharpe ratio to improve risk-adjusted returns and minimizing historical CVaR to limit losses in the worst 5% of monthly observations. The analysis considers three historical estimation windows of 3\, 5\, and 10 years and five candidate budgets ranging from 128 to 2048. We evaluate each configuration through 50 independent randomized repetitions and compare it with deterministic numerical benchmarks using approximation errors\, portfolio characteristics\, and execution times. This computational framework enables systematic analysis of how sampling techniques\, risk objectives and historical periods influence portfolio selection. Scientific impact Methodology and Computational Framework The research presents a reproducible Python-based framework for stochastic portfolio optimization\, implemented on the HEMUS supercomputer at IICT-BAS. The study examines 32 global financial assets across 3-\, 5- and 10-year historical windows\, comparing Monte Carlo (with pseudorandom generator Mersenne Twister) and randomized Quasi-Monte Carlo (with scrambled Sobol and Halton) methods under Sharpe ratio maximization and CVaR minimization. Experiment 1 – Accuracy and Convergence of Stochastic Portfolio Search: Five sampling budgets (128–2048 candidate portfolios) and 50 independent repetitions per configuration were evaluated against deterministic numerical benchmarks. Increasing the number of candidates reduced approximation errors\, while differences between sampling methods remained modest\, with classical Monte Carlo generally requiring less execution time.Experiment 2 – Financial Risk – Return Comparison: The 5-year Sharpe-oriented portfolios achieved a mean annualized return of 9.15%\, while the 10-year CVaR-oriented portfolios recorded the lowest mean historical monthly CVaR of 4.40% and annualized volatility of 6.59%. These results highlight the trade-off between maximizing risk-adjusted returns and minimizing downside risk.Experiment 3 – Historical Window and Sampling Sensitivity: The 3-\, 5- and 10-year estimation windows produced different portfolio allocations and risk–return characteristics. The optimization objective and historical period generally had a greater influence on portfolio selection than the choice of sampling generator.Scientific Contribution The study demonstrates the potential of HPC-supported stochastic optimization for reproducible financial risk analysis and numerical benchmarking. The HEMUS supercomputer enabled extensive computational experiments\, providing a foundation for scalable portfolio optimization and future financial decision-support systems. Benefits:  Improved optimization accuracy: Larger sampling budgets reduce approximation errors relative to deterministic benchmarks.Risk-aware portfolio selection: Compares return-oriented Sharpe optimization with downside-risk-focused CVaR.Reliable numerical benchmarking: Evaluates accuracy\, stability and computational efficiency across MC/RQMC methods.Better understanding of market sensitivity: Reveals how historical estimation windows influence investment outcomes.HPC-enabled research: Supports repeated simulations and scalable numerical experimentation.Future financial applications: Provides a foundation for computational portfolio analysis and investment decision-support systems.\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Success story # Highlights: Keywords: HPC; Monte Carlo; Quasi-Monte Carlo; Portfolio Optimization; Sharpe Ratio; CVaR; Financial Risk.Technology: HEMUS Supercomputer; Python; Mersenne Twister; Scrambled Sobol and Halton; Stochastic Optimization.Industry sectors: Finance/Insurance; IT/HPC Systems\, Services & Software Providers.Scientific achievement: Development of a reproducible HPC-based framework for comparing stochastic portfolio optimization methods.Key results: Larger sampling budgets improved numerical accuracy\, while optimization objectives and historical windows significantly influenced portfolio risk–return characteristics.HPC contribution: HEMUS enabled repeated simulations\, algorithm benchmarking and computational performance analysis.\n \n \n \n \n Comparison of the 3 time horizons through Sharpe-ratio optimization. Among the sampled candidates\, the 5-year Sharpe solution provides the highest expected return and the strongest reward-to-risk trade-off\, at the cost of the highest volatility across the three horizons \n \n \n \n \n Figure 1. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 2. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 3. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Comparison of the 3 time horizons through CVaR (5%). At N=100\, Mean-CVaR shifts the selected portfolios toward lower-volatility regions. The 5-year solution achieves the highest expected return while slightly reducing volatility relative to the 3-year case\, whereas the 10-year solution is the most conservative in volatility terms. \n \n \n \n \n Figure 4. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 5. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 6. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 7. Sharpe vs CVaR: Cross-Horizon Comparison. Sharpe is higher at all three horizons. CVaR is lower at all three horizons. \n \n \n \n \n \n \n \n \n \n \n Contact: Silvi-Maria Gurova | IICT-BAS | silvimaria.gurova@iict.bas.bgTodor Gurov | IICT-BAS | todor.gurov@iict.bas.bgAleksandar Kirilov | IICT-BAS | alex.kirilov@iict.bas.bg URL:https://eurocc-bulgaria.bg/?tribe_events=training-on-hpc-fundamentals-with-cybersecurity-applications-plovdiv-2-10-2025 END:VEVENT BEGIN:VEVENT DTSTART;TZID=Europe/Helsinki:20250926T080000 DTEND;TZID=Europe/Helsinki:20250928T170000 DTSTAMP:20261023T101421 CREATED:20250916T090022Z LAST-MODIFIED:20251112T134613Z UID:6595-1758873600-1759078800@eurocc-bulgaria.bg SUMMARY:International conference DiPP2025\, Burgas\, 26-28 September 2025 DESCRIPTION:NCC BULGARIA\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n NCC presenting the success story: NCC Bulgaria is the National Competence Centre for HPC and HPDA\, operating within the EuroCC3 network and EuroHPC ecosystem. Coordinated by the Institute of Information and Communication Technologies at the Bulgarian Academy of Sciences (IICT-BAS) in partnership with the Sofia University “St. Kliment Ohridski” (SU) and the University of National and World Economy (UNWE)\, it supports industry\, academia\, and public organizations through specialized training\, consulting\, technical expertise\, and access to national and European HPC resources. Scientific partners involved The Institute of Information and Communication Technologies (IICT-BAS) conducts fundamental and applied scientific research in the fields of high-performance computing (HPC)\, machine learning\, big data processing\, cybersecurity\, optimization\, financial modeling and mathematical modeling. The developed methods and technologies find application in medicine\, ecology\, industry\, transport\, financial analysis\, process optimization\, and digital transformation. Technical/scientific Challenge Constructing reliable investment portfolios requires balancing expected returns\, market volatility\, and potential losses under adverse market conditions. For a diversified universe of 32 global financial assets\, the large number of possible allocations makes constrained portfolio optimization computationally demanding. An additional challenge is that different risk measures\, such as the Sharpe ratio and Conditional Value-at-Risk (CVaR)\, may favour substantially different investment strategies. Portfolio decisions are also sensitive to the historical estimation period and the numerical sampling method. The research therefore investigates how Monte Carlo (MC) and randomized Quasi-Monte Carlo (QMC) algorithms can efficiently identify feasible portfolios while assessing numerical accuracy\, computational cost\, and sensitivity to market conditions. \n \n \n \n \n \n \n \n \n \n \n Solution Researchers at the Institute of Information and Communication Technologies of the Bulgarian Academy of Sciences (IICT-BAS) developed a reproducible Python-based framework for stochastic portfolio optimization using the HEMUS supercomputer. The methodology compares classical Monte Carlo sampling with Mersenne Twister and randomized Quasi-Monte Carlo methods based on scrambled Sobol and randomized Halton sequences. Candidate portfolios are generated under full-investment and long-only constraints\, with a maximum allocation of 30% per asset. We investigate two optimization objectives: maximizing the Sharpe ratio to improve risk-adjusted returns and minimizing historical CVaR to limit losses in the worst 5% of monthly observations. The analysis considers three historical estimation windows of 3\, 5\, and 10 years and five candidate budgets ranging from 128 to 2048. We evaluate each configuration through 50 independent randomized repetitions and compare it with deterministic numerical benchmarks using approximation errors\, portfolio characteristics\, and execution times. This computational framework enables systematic analysis of how sampling techniques\, risk objectives and historical periods influence portfolio selection. Scientific impact Methodology and Computational Framework The research presents a reproducible Python-based framework for stochastic portfolio optimization\, implemented on the HEMUS supercomputer at IICT-BAS. The study examines 32 global financial assets across 3-\, 5- and 10-year historical windows\, comparing Monte Carlo (with pseudorandom generator Mersenne Twister) and randomized Quasi-Monte Carlo (with scrambled Sobol and Halton) methods under Sharpe ratio maximization and CVaR minimization. Experiment 1 – Accuracy and Convergence of Stochastic Portfolio Search: Five sampling budgets (128–2048 candidate portfolios) and 50 independent repetitions per configuration were evaluated against deterministic numerical benchmarks. Increasing the number of candidates reduced approximation errors\, while differences between sampling methods remained modest\, with classical Monte Carlo generally requiring less execution time.Experiment 2 – Financial Risk – Return Comparison: The 5-year Sharpe-oriented portfolios achieved a mean annualized return of 9.15%\, while the 10-year CVaR-oriented portfolios recorded the lowest mean historical monthly CVaR of 4.40% and annualized volatility of 6.59%. These results highlight the trade-off between maximizing risk-adjusted returns and minimizing downside risk.Experiment 3 – Historical Window and Sampling Sensitivity: The 3-\, 5- and 10-year estimation windows produced different portfolio allocations and risk–return characteristics. The optimization objective and historical period generally had a greater influence on portfolio selection than the choice of sampling generator.Scientific Contribution The study demonstrates the potential of HPC-supported stochastic optimization for reproducible financial risk analysis and numerical benchmarking. The HEMUS supercomputer enabled extensive computational experiments\, providing a foundation for scalable portfolio optimization and future financial decision-support systems. Benefits:  Improved optimization accuracy: Larger sampling budgets reduce approximation errors relative to deterministic benchmarks.Risk-aware portfolio selection: Compares return-oriented Sharpe optimization with downside-risk-focused CVaR.Reliable numerical benchmarking: Evaluates accuracy\, stability and computational efficiency across MC/RQMC methods.Better understanding of market sensitivity: Reveals how historical estimation windows influence investment outcomes.HPC-enabled research: Supports repeated simulations and scalable numerical experimentation.Future financial applications: Provides a foundation for computational portfolio analysis and investment decision-support systems.\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Success story # Highlights: Keywords: HPC; Monte Carlo; Quasi-Monte Carlo; Portfolio Optimization; Sharpe Ratio; CVaR; Financial Risk.Technology: HEMUS Supercomputer; Python; Mersenne Twister; Scrambled Sobol and Halton; Stochastic Optimization.Industry sectors: Finance/Insurance; IT/HPC Systems\, Services & Software Providers.Scientific achievement: Development of a reproducible HPC-based framework for comparing stochastic portfolio optimization methods.Key results: Larger sampling budgets improved numerical accuracy\, while optimization objectives and historical windows significantly influenced portfolio risk–return characteristics.HPC contribution: HEMUS enabled repeated simulations\, algorithm benchmarking and computational performance analysis.\n \n \n \n \n Comparison of the 3 time horizons through Sharpe-ratio optimization. Among the sampled candidates\, the 5-year Sharpe solution provides the highest expected return and the strongest reward-to-risk trade-off\, at the cost of the highest volatility across the three horizons \n \n \n \n \n Figure 1. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 2. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 3. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Comparison of the 3 time horizons through CVaR (5%). At N=100\, Mean-CVaR shifts the selected portfolios toward lower-volatility regions. The 5-year solution achieves the highest expected return while slightly reducing volatility relative to the 3-year case\, whereas the 10-year solution is the most conservative in volatility terms. \n \n \n \n \n Figure 4. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 5. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 6. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 7. Sharpe vs CVaR: Cross-Horizon Comparison. Sharpe is higher at all three horizons. CVaR is lower at all three horizons. \n \n \n \n \n \n \n \n \n \n \n Contact: Silvi-Maria Gurova | IICT-BAS | silvimaria.gurova@iict.bas.bgTodor Gurov | IICT-BAS | todor.gurov@iict.bas.bgAleksandar Kirilov | IICT-BAS | alex.kirilov@iict.bas.bg URL:https://eurocc-bulgaria.bg/?tribe_events=international-conference-dipp2025-burgas-26-28-september-2025 END:VEVENT BEGIN:VEVENT DTSTART;TZID=Europe/Helsinki:20250925T140000 DTEND;TZID=Europe/Helsinki:20250925T150000 DTSTAMP:20261023T101421 CREATED:20250910T090010Z LAST-MODIFIED:20260423T112002Z UID:6580-1758808800-1758812400@eurocc-bulgaria.bg SUMMARY:Seminar:"Organization of the processing of large volumes of economic data on Avitohol"\, 25.09.2025 DESCRIPTION:NCC BULGARIA\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n NCC presenting the success story: NCC Bulgaria is the National Competence Centre for HPC and HPDA\, operating within the EuroCC3 network and EuroHPC ecosystem. Coordinated by the Institute of Information and Communication Technologies at the Bulgarian Academy of Sciences (IICT-BAS) in partnership with the Sofia University “St. Kliment Ohridski” (SU) and the University of National and World Economy (UNWE)\, it supports industry\, academia\, and public organizations through specialized training\, consulting\, technical expertise\, and access to national and European HPC resources. Scientific partners involved The Institute of Information and Communication Technologies (IICT-BAS) conducts fundamental and applied scientific research in the fields of high-performance computing (HPC)\, machine learning\, big data processing\, cybersecurity\, optimization\, financial modeling and mathematical modeling. The developed methods and technologies find application in medicine\, ecology\, industry\, transport\, financial analysis\, process optimization\, and digital transformation. Technical/scientific Challenge Constructing reliable investment portfolios requires balancing expected returns\, market volatility\, and potential losses under adverse market conditions. For a diversified universe of 32 global financial assets\, the large number of possible allocations makes constrained portfolio optimization computationally demanding. An additional challenge is that different risk measures\, such as the Sharpe ratio and Conditional Value-at-Risk (CVaR)\, may favour substantially different investment strategies. Portfolio decisions are also sensitive to the historical estimation period and the numerical sampling method. The research therefore investigates how Monte Carlo (MC) and randomized Quasi-Monte Carlo (QMC) algorithms can efficiently identify feasible portfolios while assessing numerical accuracy\, computational cost\, and sensitivity to market conditions. \n \n \n \n \n \n \n \n \n \n \n Solution Researchers at the Institute of Information and Communication Technologies of the Bulgarian Academy of Sciences (IICT-BAS) developed a reproducible Python-based framework for stochastic portfolio optimization using the HEMUS supercomputer. The methodology compares classical Monte Carlo sampling with Mersenne Twister and randomized Quasi-Monte Carlo methods based on scrambled Sobol and randomized Halton sequences. Candidate portfolios are generated under full-investment and long-only constraints\, with a maximum allocation of 30% per asset. We investigate two optimization objectives: maximizing the Sharpe ratio to improve risk-adjusted returns and minimizing historical CVaR to limit losses in the worst 5% of monthly observations. The analysis considers three historical estimation windows of 3\, 5\, and 10 years and five candidate budgets ranging from 128 to 2048. We evaluate each configuration through 50 independent randomized repetitions and compare it with deterministic numerical benchmarks using approximation errors\, portfolio characteristics\, and execution times. This computational framework enables systematic analysis of how sampling techniques\, risk objectives and historical periods influence portfolio selection. Scientific impact Methodology and Computational Framework The research presents a reproducible Python-based framework for stochastic portfolio optimization\, implemented on the HEMUS supercomputer at IICT-BAS. The study examines 32 global financial assets across 3-\, 5- and 10-year historical windows\, comparing Monte Carlo (with pseudorandom generator Mersenne Twister) and randomized Quasi-Monte Carlo (with scrambled Sobol and Halton) methods under Sharpe ratio maximization and CVaR minimization. Experiment 1 – Accuracy and Convergence of Stochastic Portfolio Search: Five sampling budgets (128–2048 candidate portfolios) and 50 independent repetitions per configuration were evaluated against deterministic numerical benchmarks. Increasing the number of candidates reduced approximation errors\, while differences between sampling methods remained modest\, with classical Monte Carlo generally requiring less execution time.Experiment 2 – Financial Risk – Return Comparison: The 5-year Sharpe-oriented portfolios achieved a mean annualized return of 9.15%\, while the 10-year CVaR-oriented portfolios recorded the lowest mean historical monthly CVaR of 4.40% and annualized volatility of 6.59%. These results highlight the trade-off between maximizing risk-adjusted returns and minimizing downside risk.Experiment 3 – Historical Window and Sampling Sensitivity: The 3-\, 5- and 10-year estimation windows produced different portfolio allocations and risk–return characteristics. The optimization objective and historical period generally had a greater influence on portfolio selection than the choice of sampling generator.Scientific Contribution The study demonstrates the potential of HPC-supported stochastic optimization for reproducible financial risk analysis and numerical benchmarking. The HEMUS supercomputer enabled extensive computational experiments\, providing a foundation for scalable portfolio optimization and future financial decision-support systems. Benefits:  Improved optimization accuracy: Larger sampling budgets reduce approximation errors relative to deterministic benchmarks.Risk-aware portfolio selection: Compares return-oriented Sharpe optimization with downside-risk-focused CVaR.Reliable numerical benchmarking: Evaluates accuracy\, stability and computational efficiency across MC/RQMC methods.Better understanding of market sensitivity: Reveals how historical estimation windows influence investment outcomes.HPC-enabled research: Supports repeated simulations and scalable numerical experimentation.Future financial applications: Provides a foundation for computational portfolio analysis and investment decision-support systems.\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Success story # Highlights: Keywords: HPC; Monte Carlo; Quasi-Monte Carlo; Portfolio Optimization; Sharpe Ratio; CVaR; Financial Risk.Technology: HEMUS Supercomputer; Python; Mersenne Twister; Scrambled Sobol and Halton; Stochastic Optimization.Industry sectors: Finance/Insurance; IT/HPC Systems\, Services & Software Providers.Scientific achievement: Development of a reproducible HPC-based framework for comparing stochastic portfolio optimization methods.Key results: Larger sampling budgets improved numerical accuracy\, while optimization objectives and historical windows significantly influenced portfolio risk–return characteristics.HPC contribution: HEMUS enabled repeated simulations\, algorithm benchmarking and computational performance analysis.\n \n \n \n \n Comparison of the 3 time horizons through Sharpe-ratio optimization. Among the sampled candidates\, the 5-year Sharpe solution provides the highest expected return and the strongest reward-to-risk trade-off\, at the cost of the highest volatility across the three horizons \n \n \n \n \n Figure 1. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 2. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 3. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n Comparison of the 3 time horizons through CVaR (5%). At N=100\, Mean-CVaR shifts the selected portfolios toward lower-volatility regions. The 5-year solution achieves the highest expected return while slightly reducing volatility relative to the 3-year case\, whereas the 10-year solution is the most conservative in volatility terms. \n \n \n \n \n Figure 4. 3 – year hori \n \n \n \n \n \n \n \n \n \n \n Figure 5. 5- year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 6. 10-year horizon \n \n \n \n \n \n \n \n \n \n \n Figure 7. Sharpe vs CVaR: Cross-Horizon Comparison. Sharpe is higher at all three horizons. CVaR is lower at all three horizons. \n \n \n \n \n \n \n \n \n \n \n Contact: Silvi-Maria Gurova | IICT-BAS | silvimaria.gurova@iict.bas.bgTodor Gurov | IICT-BAS | todor.gurov@iict.bas.bgAleksandar Kirilov | IICT-BAS | alex.kirilov@iict.bas.bg URL:https://eurocc-bulgaria.bg/?tribe_events=organization-of-the-processing-of-large-volumes-of-economic-data-on-avitohol-25-09-2025 END:VEVENT END:VCALENDAR