Deep Reinforcement Learning For Nash Equilibria in Non-Renewable Resource Differential Games

El Mehdi Nezahi, Loubna Benabbou, Hassan Benchekroun

Keywords: Reinforcement Learning, Game Theory, Differential Resource Games

Abstract: Characterizing Nash equilibria in oligopolistic non-renewable resource markets poses major challenges for computational economics, as traditional iterative methods face scalability limitations due to the curse of dimensionality. In this work, we propose a reinforcement learning–based approach to compute these equilibria and benchmark it against a modified iterative baseline, derived from an established algorithm for differential games and adapted to the oligopoly case. We conduct experiments in monopoly, duopoly, and multi-player settings, evaluating both reward accuracy and computational efficiency. Our results show that while iterative schemes provide good accuracy in low-dimensional problems, reinforcement learning scales more effectively to three- and four-player games, leading to a substantial reduction in computation time. This highlights the potential of reinforcement learning as a scalable tool for solving complex differential games in resource economics.

Primary Area: reinforcement learning

Submission Number: 14268