Deep Q-Learning for Nash Equilibria: Nash-DQN
This addresses the challenge of applying reinforcement learning to general-sum games beyond zero-sum or simplified settings, which is incremental as it builds on existing methods with a novel expansion approach.
The paper tackled the problem of model-free learning for Nash equilibria in general-sum stochastic games, developing a data-efficient Deep Q-learning method that uses a local linear-quadratic expansion parameterized by neural networks, and applied it to learning optimal trading strategies in competitive electronic markets as a proof of concept.
Model-free learning for multi-agent stochastic games is an active area of research. Existing reinforcement learning algorithms, however, are often restricted to zero-sum games, and are applicable only in small state-action spaces or other simplified settings. Here, we develop a new data efficient Deep-Q-learning methodology for model-free learning of Nash equilibria for general-sum stochastic games. The algorithm uses a local linear-quadratic expansion of the stochastic game, which leads to analytically solvable optimal actions. The expansion is parametrized by deep neural networks to give it sufficient flexibility to learn the environment without the need to experience all state-action pairs. We study symmetry properties of the algorithm stemming from label-invariant stochastic games and as a proof of concept, apply our algorithm to learning optimal trading strategies in competitive electronic markets.