Multi-Agent Reinforcement Learning via Double Averaging Primal-Dual Optimization
This addresses the challenge of complex interactions in MARL for applications like sensor networks and swarm robotics, though it appears incremental as it builds on existing primal-dual methods.
The paper tackles policy evaluation in multi-agent reinforcement learning by proposing a double averaging primal-dual optimization algorithm, which converges to the optimal solution at a global geometric rate, achieving fast finite-time convergence for decentralized convex-concave saddle-point problems.
Despite the success of single-agent reinforcement learning, multi-agent reinforcement learning (MARL) remains challenging due to complex interactions between agents. Motivated by decentralized applications such as sensor networks, swarm robotics, and power grids, we study policy evaluation in MARL, where agents with jointly observed state-action pairs and private local rewards collaborate to learn the value of a given policy. In this paper, we propose a double averaging scheme, where each agent iteratively performs averaging over both space and time to incorporate neighboring gradient information and local reward information, respectively. We prove that the proposed algorithm converges to the optimal solution at a global geometric rate. In particular, such an algorithm is built upon a primal-dual reformulation of the mean squared projected Bellman error minimization problem, which gives rise to a decentralized convex-concave saddle-point problem. To the best of our knowledge, the proposed double averaging primal-dual optimization algorithm is the first to achieve fast finite-time convergence on decentralized convex-concave saddle-point problems.