GTLGMLJun 17, 2020

Policy Evaluation and Seeking for Multi-Agent Reinforcement Learning via Best Response

arXiv:2006.09585v21 citations
AI Analysis

This work addresses policy evaluation challenges in multi-agent reinforcement learning, offering a novel approach for handling dynamical behaviors, but it is incremental as it builds on existing game-theoretic concepts like sink equilibrium.

The paper tackles the problem of evaluating and ranking policies in multi-agent reinforcement learning by introducing cycle-based and memory-based metrics based on sink equilibrium, using strict best response dynamics to handle cyclical behaviors and non-stationarity, with results showing that policies with maximum metric are observed with nonzero probability or within a given tolerance of optimal.

This paper introduces two metrics (cycle-based and memory-based metrics), grounded on a dynamical game-theoretic solution concept called sink equilibrium, for the evaluation, ranking, and computation of policies in multi-agent learning. We adopt strict best response dynamics (SBRD) to model selfish behaviors at a meta-level for multi-agent reinforcement learning. Our approach can deal with dynamical cyclical behaviors (unlike approaches based on Nash equilibria and Elo ratings), and is more compatible with single-agent reinforcement learning than alpha-rank which relies on weakly better responses. We first consider settings where the difference between largest and second largest underlying metric has a known lower bound. With this knowledge we propose a class of perturbed SBRD with the following property: only policies with maximum metric are observed with nonzero probability for a broad class of stochastic games with finite memory. We then consider settings where the lower bound for the difference is unknown. For this setting, we propose a class of perturbed SBRD such that the metrics of the policies observed with nonzero probability differ from the optimal by any given tolerance. The proposed perturbed SBRD addresses the opponent-induced non-stationarity by fixing the strategies of others for the learning agent, and uses empirical game-theoretic analysis to estimate payoffs for each strategy profile obtained due to the perturbation.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes