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

Coopetition-Gym v1 is a formally grounded benchmark platform for multi-agent reinforcement learning (MARL) under strategic coopetition, introduced by researchers Vik Pant and Eric Yu. The platform comprises twenty environments across four mechanism classes derived from foundational technical reports: interdependence and complementarity, trust and reputation dynamics, collective action and loyalty, and sequential interaction and reciprocity. Each environment features a closed-form payoff structure and a calibrated interdependence matrix, with a parameterized reward layer configurable in three modes (private, integrated, cooperative) to enable reward-type ablation. Notably, four environments are validated against historically documented coopetitive relationships, reproducing outcomes at 98.3% (Samsung-Sony LCD), 81.7% (Renault-Nissan Alliance), 86.7% (Apache HTTP Server), and 87.3% (Apple iOS App Store). The platform exposes Gymnasium, PettingZoo Parallel, and PettingZoo AEC interfaces, and ships 126 reference algorithms including 16 learning algorithms, 7 game-theoretic oracles, 2 heuristic baselines, and 101 constant-action policies.

The significance of Coopetition-Gym v1 lies in its formal grounding and comprehensive tooling for mixed-motive scenarios where agents must both cooperate and compete. 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. It 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. This enables researchers to systematically study reward functions, trust dynamics, and strategic interactions in multi-agent systems—critical for applications in autonomous negotiation, supply chain management, and collaborative robotics.