Phi-Actor-Critic: Steering General-Sum Games to Pareto-Efficient Correlated Equilibria

June 09, 2026 Β· Grace Period Β· πŸ› IJCAI 2026

⏳ Grace Period
This paper is less than 90 days old. We give authors time to release their code before passing judgment.
Authors Wongyu Lee, Francesco Lelli, Omran Ayoub, Massimo Tornatore arXiv ID 2606.11284 Category cs.MA: Multiagent Systems Cross-listed cs.GT, cs.LG Citations 0 Venue IJCAI 2026
Abstract
Real-world multi-agent systems, from traffic coordination to resource allocation, are often modeled as general-sum games where individual incentives conflict with collective welfare. In these settings, the central challenge is not merely finding an equilibrium, but selecting socially desirable outcomes among many suboptimal Nash equilibria. Standard deep multi-agent reinforcement learning (MARL) methods struggle with this problem, as value-decomposition approaches are constrained by monotonicity assumptions and policy-gradient methods often converge to stable but socially inefficient equilibria. To address this limitation, we propose $Ξ¦$-Actor-Critic ($Ξ¦$-AC), a framework that leverages swap regret minimization to steer learning toward high-welfare correlated equilibria (CE). To make counterfactual regret estimation tractable in deep MARL, $Ξ¦$-AC employs a centralized attention critic that predicts vector-valued regrets in a single forward pass, avoiding computationally expensive counterfactual simulations. We further introduce a Lagrangian-based equilibrium selection mechanism that optimizes social welfare while enforcing stability through regret constraints. Experiments on matrix games, Multi-Agent Particle Environments (MPE), and the Melting Pot Harvest scenario demonstrate that $Ξ¦$-AC learns efficient and stable coordination strategies across diverse mixed-motive settings while maintaining high collective return and competitive fairness.
Community shame:
Not yet rated
Community Contributions

Found the code? Know the venue? Think something is wrong? Let us know!

πŸ“œ Similar Papers

In the same crypt β€” Multiagent Systems