Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Zhang, Junyu, Yang, Feihong, Wang, Jian, Wang, Chao, Zhang, Xudong
Format: Preprint
Veröffentlicht: 2026
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2605.28273
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866917540189437952
author Zhang, Junyu
Yang, Feihong
Wang, Jian
Wang, Chao
Zhang, Xudong
author_facet Zhang, Junyu
Yang, Feihong
Wang, Jian
Wang, Chao
Zhang, Xudong
contents The Policy-Space Response Oracles (PSRO) framework scales equilibrium computation to large zero-sum games by iteratively expanding a restricted strategy set using deep reinforcement learning (DRL). A central challenge is to construct, under limited computational budgets, a small strategy population whose induced game well approximates the full game. Existing PSRO variants typically expand the population using best responses to meta-strategies computed from restricted-game payoffs, which can lead to inefficient expansions that provide limited global improvement. We propose to guide population expansion by directly evaluating the post-expansion population quality. Specifically, we adopt Population Exploitability (PE) to measure how well a restricted strategy set represents the full game, and introduce a two-phase exploration--selection framework that explicitly minimizes PE during expansion. We instantiate this framework as Global PSRO, a practical DRL-based algorithm that efficiently generates candidate responses and estimates PE via parameter-sharing conditional neural networks. Experiments across multiple two-player zero-sum games show that Global PSRO achieves lower exploitability and approximates Nash equilibria with significantly fewer policy iterations than prior PSRO methods.
format Preprint
id arxiv_https___arxiv_org_abs_2605_28273
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Global Policy-Space Response Oracles for Two-Player Zero-Sum Games
Zhang, Junyu
Yang, Feihong
Wang, Jian
Wang, Chao
Zhang, Xudong
Artificial Intelligence
The Policy-Space Response Oracles (PSRO) framework scales equilibrium computation to large zero-sum games by iteratively expanding a restricted strategy set using deep reinforcement learning (DRL). A central challenge is to construct, under limited computational budgets, a small strategy population whose induced game well approximates the full game. Existing PSRO variants typically expand the population using best responses to meta-strategies computed from restricted-game payoffs, which can lead to inefficient expansions that provide limited global improvement. We propose to guide population expansion by directly evaluating the post-expansion population quality. Specifically, we adopt Population Exploitability (PE) to measure how well a restricted strategy set represents the full game, and introduce a two-phase exploration--selection framework that explicitly minimizes PE during expansion. We instantiate this framework as Global PSRO, a practical DRL-based algorithm that efficiently generates candidate responses and estimates PE via parameter-sharing conditional neural networks. Experiments across multiple two-player zero-sum games show that Global PSRO achieves lower exploitability and approximates Nash equilibria with significantly fewer policy iterations than prior PSRO methods.
title Global Policy-Space Response Oracles for Two-Player Zero-Sum Games
topic Artificial Intelligence
url https://arxiv.org/abs/2605.28273