Coopetition-Gym v1: A Formally Grounded Platform for Mixed-Motive Multi-Agent Reinforcement Learning under Strategic Coopetition

2026-05-03

AI Summary

This work addresses the lack of standardized benchmarks for multi-agent reinforcement learning in strategic coopetition settings, where agents exhibit mixed motives of cooperation and competition. The authors propose the first formal benchmark platform, comprising 20 environments derived from four distinct mechanism designs, each featuring closed-form payoff functions, calibrated interdependence matrices, and support for three reward schemes to facilitate ablation studies. The platform innovatively integrates continuous action spaces, parameterized reward reciprocity, game-theoretic Oracle baselines, and validation against historical coopetition cases. It is compatible with Gymnasium and PettingZoo interfaces, incorporates 126 algorithms, and releases open-source data from 25,708 training runs and 1,116 behavioral audits. Reproduction of four historical cases achieves success rates of 98.3%, 81.7%, 86.7%, and 87.3%, respectively.

Abstract

We present Coopetition-Gym v1, a benchmark platform for mixed-motive multi-agent reinforcement learning under strategic coopetition. The platform comprises twenty environments organized into four mechanism classes that correspond to four foundational technical reports: interdependence and complementarity (arXiv:2510.18802), trust and reputation dynamics (arXiv:2510.24909), collective action and loyalty (arXiv:2601.16237), and sequential interaction and reciprocity (arXiv:2604.01240). Each environment carries a closed-form payoff structure and a calibrated interdependence matrix derived from the corresponding report. Every environment exposes a parameterized reward layer configurable across three structurally distinct modes (private, integrated, cooperative). This separation of payoff from reward enables reward-type ablation, the platform's principal methodological apparatus. Four of the twenty environments are calibrated against historically documented coopetitive relationships and reproduce their outcomes at 98.3, 81.7, 86.7, and 87.3 percent on the validation rubric (Samsung-Sony LCD, Renault-Nissan Alliance, Apache HTTP Server, Apple iOS App Store). The platform exposes Gymnasium, PettingZoo Parallel, and PettingZoo AEC interfaces and ships 126 reference algorithms: 16 learning algorithms, 7 game-theoretic oracles, 2 heuristic baselines, and 101 constant-action policies. A reference experimental study trained the 16 learning algorithms on every environment under every reward configuration with seven random seeds, producing a 25,708-run training corpus and a 1,116-run behavioral audit corpus, both released under CC-BY-4.0 with Croissant 1.0 metadata. Coopetition-Gym v1 is the first platform to combine continuous-action mixed-motive environments, parameterized reward mutuality, calibrated interdependence coefficients, game-theoretic oracle baselines, and validated case studies.

Problem

• coopetition
• mixed-motive
• multi-agent reinforcement learning
• strategic interaction
• interdependence

Innovation

• coopetition
• multi-agent reinforcement learning
• mixed-motive games
• reward ablation
• interdependence calibration