uk

Article

Machine learning algorithms for process modelling and decision-making in project portfolio management

Yurii Yushchenko
Abstract

Project portfolio management in dynamic and uncertain environments increasingly requires methods capable of supporting rapid decision-making, continuous adaptation, and resilience against external volatility. Recent advances in machine learning provide a foundation for integrating algorithmic intelligence into portfolio-level processes, enabling organisations to select, prioritise, and adjust project configurations in real time. The purpose of this article was to develop and formalise an intelligent framework for adaptive project portfolio management based on the mathematical foundations of dynamic reinforcement learning algorithms. To achieve this goal, a complex of methods was applied, including mathematical modelling of decision-making processes using Multi-Armed Bandits, synthesis of the Upper Confidence Bound algorithm family, and scenario-based simulation for a comparative analysis of the proposed approaches’ effectiveness. The central result of the study was the justification of the advantages of the Dynamic Confidence Bound algorithm, which, through an exponential discounting mechanism, allowed the system to disregard outdated data and focus on current performance indicators. Experimental validation established that the use of machine learning increases cumulative reward by 18-22% compared to heuristic methods in stable environments, while in non-stationary conditions, Dynamic Confidence Bound outperforms classical approaches by 14-17%. Simulation results confirmed that the proposed model detects project performance degradation or shifts 2 to 4 times faster than standard mechanisms, minimising cognitive biases, particularly anchoring. It has been demonstrated that the implementation of adaptive discounting ensures 48-60% faster portfolio recovery after sharp external shocks compared to base Upper Confidence Bound algorithms. The study also demonstrated high model sensitivity to hyperparameter tuning, allowing for a flexible balance between the exploration of new opportunities and the exploitation of proven solutions depending on the organisation’s strategic context. The practical significance of the work lies in the creation of a ready-to-use computational pipeline that can be integrated into corporate project management systems to automate prioritisation and dynamic resource reallocation in real time

Keywords

adaptive decision-making; Multi-Armed Bandit; Upper Confidence Bound; dynamic environments

Download article

Received 10.10.2025, Revised 29.01.2026, Accepted 24.02.2026 Published 21.05.2026

Retrieved from Vol. 13, No. 1, 2026

Suggested citation

Yushchenko, Yu. (2026). Machine learning algorithms for process modelling and decision-making in project portfolio management. Economics, Entrepreneurship, Management, 13(1), 76-84. https://doi.org/10.56318/ eem2026.01.076

https://doi.org/10.56318/ eem2026.01.076

Pages 76-84

References

  1. Bertsimas, D., & Kallus, N. (2018). From predictive to prescriptive analytics. Management Science, 66(3), 1025-1044. doi: 10.1287/mnsc.2018.3253.
  2. Bondar, O., Ryzhenko, N., & Chernykhivska, A. (2023). Green recovery of Ukraine in the post-war period. Collection of Scientific Papers “ΛΌГOΣ”, 13-14. doi: 10.36074/logos-27.10.2023.01.
  3. Bubeck, S., & Cesa-Bianchi, N. (2012). Regret analysis of stochastic and non-stochastic multi-armed bandit problems. Foundations and Trends in Machine Learning, 5(1), 1-122. doi: 10.1561/9781601986276.
  4. Bushuyev, S., Bushuyeva, N., Lobok, Ye., & Murovansky, H. (2025). Management of innovative projects based on artificial intelligence applications in a turbulent environment. Innovative Technologies and Scientific Solutions for Industries, 2(32), 168-176. doi: 10.30837/2522-9818.2025.2.168.
  5. Bushuyev, S., Bushuyeva, N., Onyshchenko, S., & Bondar, A. (2021). Modelling projects portfolio structure dynamics. In 2021 IEEE 16th international conference on computer sciences and information technologies (pp. 293-298). Lviv: IEEE Ukraine Section. doi: 10.1109/CSIT52700.2021.9648713.
  6. Chakraborty, S. (2022). Incentivized exploration of non-stationary stochastic bandits. (Master's thesis, University of Colorado, Boulder, USA).
  7. Chen, Q., Golrezaei, N., & Bouneffouf, D. (2022). Non-stationary bandits with auto-regressive temporal dependency. In Proceedings of the 37th international conference on neural information processing systems (pp. 7895-7929). New York: Curran Associates Inc. doi: 10.5555/3666122.3666468.
  8. Cooper, R.G. (2022). The 5th generation stage-gate idea to launch process. IEEE Engineering Management Review, 50(4), 43-55. doi: 10.1109/EMR.2022.3222937.
  9. Devlin, J., Chang, M.W., Lee, K., & Toutanova, K. (2019). BERT: Pre-training of deep bidirectional transformers. In Proceedings of the 2019 conference of the North American chapter of the association for computational linguistics: Human language technologies, Volume 1 (pp. 4171-4186). Minneapolis: Association for Computational Linguistics.
  10. Esmaeili, S.A., Shin, S., & Slivkins, A. (2023). Robust and performance incentivizing algorithms for multi-armed bandits with strategic agents. Proceedings of the AAAI Conference on Artificial Intelligence, 39(13), 13814-13822. doi: 10.1609/aaai.v39i13.33510.
  11. Garivier, A., & Moulines, E. (2011). On Upper-Confidence Bound policies for Switching Bandit problems. In J. Kivinen, C. Szepesvári, E. Ukkonen & T. Zeugmann (Eds.), Algorithmic learning theory (pp. 174-188). Berlin: Springer. doi: 10.1007/978-3-642-24412-4_16.
  12. Gornet, J., Hosseinzadeh, M., & Sinopoli, B. (2022). Stochastic Multi-armed Bandits with non-stationary rewards generated by a linear dynamical system. In IEEE 61st conference on decision and control (pp. 1460-1465). Mexico: Institute of Electrical and Electronics Engineers.
  13. Harris, K., & Slivkins, A. (2025). Should you use your large language model to explore or exploit? arXiv. doi: 10.48550/arXiv.2502.00225.
  14. Keplinger, N.S., Luo, B., Bektas, I., Zhang, Y., Wray, K.H., Laszka, A., Dubey, A., & Mukhopadhyay, A. (2025). NS-Gym: A сomprehensive and open-source simulation environments and benchmarks for non-stationary Markov decision processes. arXiv. doi: 10.48550/arXiv.2501.09646.
  15. Khurshid, S., Abdulla, M.S., & Ghatak, G. (2024). Optimizing sharpe ratio: Risk-adjusted decision-making in multi-armed bandits. Machine Learning, 114, article number 32. doi: 10.1007/s10994-024-06680-2.
  16. Kovari, A. (2024). AI for decision support: Balancing accuracy, transparency, and trust across sectors. Information, 15(11), article number 725. doi: 10.3390/info15110725.
  17. Levine, S., Kumar, A., Tucker, G., & Fu, J. (2020). Offline reinforcement learning: Tutorial, review, and perspectives on open problems. arXiv. doi: 10.48550/arXiv.2005.01643.
  18. Slivkins, A., Xu, Y., & Zuo, S. (2025). Greedy algorithm for structured bandits: A sharp characterization of asymptotic success / failure. arXiv. doi: 10.48550/arXiv.2503.04010.
  19. Slivkins, A. (2019). Introduction to multi-armed bandits. arXiv. doi: 10.48550/arXiv.1904.07272.
  20. Thananjeyan, B., Kandasamy, K., Stoica, I., Jordan, M.I., Goldberg, K., & Gonzalez, J.E. (2021). Resource allocation in Multi-Armed Bandit exploration: Overcoming sublinear scaling with adaptive parallelism. Proceedings of the 38th International Conference on Machine Learning, 139, 10236-10246.
  21. Vernade, С., Gyorgy, A., & Mann, T.A. (2020). Non-stationary delayed bandits with intermediate observations. Proceedings of the 38th International Conference on Machine Learning, 119, 9722-9732.
  22. Xiang, D., West, R., Wang, J., Cui, X., & Huang, J. (2022). Multi Armed Bandit vs. A/B Tests in E-commerce – confidence interval and hypothesis test power perspectives. In KDD ‘22: Proceedings of the 28th ACM SIGKDD conference on knowledge discovery and data mining (pp. 4204-4214). New York: Association for Computing Machinery. doi: 10.1145/3534678.3539144.
  23. Yasinetskyi, O., & Galchenko, I. (2025). AI tools and risk management methods in the scrum project lifecycle. Telecommunications and Information Technologies, 3(88), 106-112. doi: 10.31673/2412-4338.2025.038711.
  24. Zuo, J., & Joe-Wong, C. (2021). Combinatorial multi-armed bandits for resource allocation. In 55th annual conference on information sciences and systems (pp. 1-4). Baltimore: Johns Hopkins University.