Regularization for Strategy Exploration in Empirical Game-Theoretic Analysis

TL;DR

This paper introduces a regularized strategy exploration method called RRD for empirical game-theoretic analysis, demonstrating its advantages over existing approaches in adaptive performance and extending it to complex three-player games.

Contribution

It proposes a novel regularized MSS called RRD, providing a more adaptive strategy exploration method and extending the framework to three-player games with incomplete models.

Findings

RRD outperforms existing MSSs in various games.

Regularization significantly improves strategy exploration performance.

The regret of best response targets critically influences exploration effectiveness.

Abstract

In iterative approaches to empirical game-theoretic analysis (EGTA), the strategy space is expanded incrementally based on analysis of intermediate game models. A common approach to strategy exploration, represented by the double oracle algorithm, is to add strategies that best-respond to a current equilibrium. This approach may suffer from overfitting and other limitations, leading the developers of the policy-space response oracle (PSRO) framework for iterative EGTA to generalize the target of best response, employing what they term meta-strategy solvers (MSSs). Noting that many MSSs can be viewed as perturbed or approximated versions of Nash equilibrium, we adopt an explicit regularization perspective to the specification and analysis of MSSs. We propose a novel MSS called regularized replicator dynamics (RRD), which simply truncates the process based on a regret criterion. We show that RRD is more adaptive than existing MSSs and outperforms them in various games. We extend our study to three-player games, for which the payoff matrix is cubic in the number of strategies and so exhaustively evaluating profiles may not be feasible. We propose a profile search method that can identify solutions from incomplete models, and combine this with iterative model construction using a regularized MSS. Finally, and most importantly, we reveal that the regret of best response targets has a tremendous influence on the performance of strategy exploration through experiments, which provides an explanation for the effectiveness of regularization in PSRO.