Mathematical modeling of agent learning processes in an organizational system

The article discusses mathematical models of agent learning processes in an organizational system. Learning is understood as a change in the agent's behavior and knowledge under the influence of experience, feedback, and interaction with other agents. A classification of models is proposed: executive learning (learning to act), cognitive learning (forming mental models), and social learning (imitation, knowledge transfer). The main equations of knowledge dynamics (Rescher-Schultz models, Bayesian updating, and Q-learning in a multi-armed bandit) are formalized. The paper proposes an equation for the iterative learning curve, based on which an algorithm is developed for constructing an individual learning trajectory for the agent, which allows for determining the learning time required to achieve the optimal level of competence for the agent. This algorithm solves the maximin problem: maximizing the minimum competence level across training areas. The proposed algorithm provides an analytical solution. The article also proposes a mathematical model of the agent's individual learning trajectory, which is based on the classical learning curve equation and allows for determining the learning time in different areas of training to ensure maximum overall assessment of the agent's competence level. The solution is obtained using numerical methods. The article identifies promising areas for research on the agent's learning processes.

1. Argyris Ch. On organizational learning. Moscow: INFRA-M; 2004. 563 p. (In Russ.).

2. Pavlov O.V. Construction of regression models of learning curves for the industrial enterprises. Vestnik of Samara University. Economics and Мanagement. 2017;8(3):60–69. (In Russ.).

3. Novikov D.A. Models of learning in the process of work. Large-Scale Systems Control. 2007;(19):5–22. (In Russ.).

4. Belov M.V., Novikov D.A., Rogatkin A.D. Learning curves estimates. Large-Scale Systems Control. 2019;(82):6–27. (In Russ.).

5. Sutton R.S., Barto A.G. Reinforcement Learning: An Introduction. Cambridge: MIT Press; 2020. 552 p.

6. Argyris Ch. Overcoming Organizational Defenses: Facilitating Organizational Learning. Boston: Allyn & Bacon; 1990. 168 p.

7. Bolshakov V.E., Sakulin S.A., Alfimtsev A.N. Multi-agent reinforcement learning using the collective intrinsic motivation. Herald of the Bauman Moscow State Technical University. Series: Instrument Engineering. 2023;(4):61–85. (In Russ.).

8. Kuzmin A.M., Vysokovskaya E.A. The Porter-Lawler motivation model is one of the tools for motivating people and influencing their behavior. Methods of Quality Management. 2021;(8):46. (In Russ.).

9. Tarasov O.V. Constructing personalized educational trajectories based on bayesian knowledge estimation, forgetting curve and task clustering. The Caspian Journal: Management and High Technologies. 2025;(1):30–37. (In Russ.).

10. Nigmatulin N.R., Belyaev P.L. Mathematical modeling in economic problems. In: Current problems and trends in the development of modern economics and computer science: Proceedings of the III International Scientific and Practical Conference: Part II, 21 November 2025, Birsk, Russia. Birsk: Ufa University of Science and Technology (Birsk branch); 2025. P. 146–149. (In Russ.).

11. Anzanello M.J., Fogliatto F.S. Learning Curve Models and Applications: Literature Review and Research Directions. International Journal of Industrial Ergonomics. 2011;41(5):573–583. https://doi.org/10.1016/j.ergon.2011.05.001

12. Reznikov D.A. Mathematical modeling of the learning process by the approximation method. Vestnik Voronezhskogo instituta FSIN Rossii. 2024;(2):103–108. (In Russ.).

13. Reznikov D.A. Development of an algorithm for solving the problem of distributing training time in training areas. Modeling, Optimization and Information Technology. 2024;12(2). (In Russ.). https://doi.org/10.26102/2310-6018/2024.45.2.008

14. Reznikov D.A. Development of a mathematical model of an employee's individual physical training path. Vestnik Voronezhskogo instituta FSIN Rossii. 2026;(1):103–106. (In Russ.).