Properties of Nash Equilibria in Differential Games with Constant Delay Time

Authors

  • Vladislav Iosifovich Zhukovsky Moscow State University, Moscow, Russian Federation
  • Lidiya Vladislavovna Zhukovskaya Central Economic and Mathematical Institute of the Russian Academy of Sciences, Moscow
  • Konstantin Nikolaevich Kudryavtsev Financial University under the Government of the Russian Federation, Moscow
  • Irina Stanislavovna Stabulit Financial University under the Government of the Russian Federation, Moscow

DOI:

https://doi.org/10.14529/mmph260402

Keywords:

differential games, Nash equilibrium, Pareto optimality, time delay

Abstract

This paper studies properties of Nash equilibria in a two-person nonzero-sum differential game with constant time delay. It is shown that equilibrium strategies may be non-unique and are not always efficient from the standpoint of the joint improvement of both players' criteria. Methods for overcoming these shortcomings are specified. The study employs the method of Lyapunov–Chetaev functions, developed by N.N. Krasovskii as applied to problems of analytical design of optimal controllers in time-delay systems. Sufficient conditions are formulated for the existence of admissible strategies that simultaneously reduce the values of the payoff functionals of both players compared to their equilibrium values. For linear games with quadratic payoff functionals, these conditions are reduced to verifying the definite negativity of matrix functions depending on the solutions to systems of differential and differential-difference equations. Explicit forms of the improving strategies are obtained for a special case where the control matrices and the weight matrices of the criteria coincide with the identity matrix. In addition, a class of equilibria possessing the property of Pareto optimality is identified: for these, any joint deviation cannot simultaneously improve the criteria of both players. Sufficient conditions for Pareto optimality and semi-efficiency of equilibrium strategies in linear-quadratic time-delay games are specified. The results can be applied to designing robust strategies in multi-agent systems, economic models, and control problems with delayed feedback.

Author Biographies

Vladislav Iosifovich Zhukovsky, Moscow State University, Moscow, Russian Federation

Dr. Sc. (Physics and Mathematics), Professor, Department of Optimal Control, Faculty of Mathematics and Mechanics (VMiK)

Lidiya Vladislavovna Zhukovskaya, Central Economic and Mathematical Institute of the Russian Academy of Sciences, Moscow

Dr. Sc. (Economics), Cand. Sc. (Physics and Mathematics), Leading Researcher

Konstantin Nikolaevich Kudryavtsev, Financial University under the Government of the Russian Federation, Moscow

Cand. Sc. (Physics and Mathematics), Associate Professor, Department of Modeling and Systems Analysis

Irina Stanislavovna Stabulit, Financial University under the Government of the Russian Federation, Moscow

Senior Lecturer, Department of Modeling and Systems Analysis

Published

2026-11-13