A Structured Prediction Approach for Generalization in Cooperative Multi-Agent Reinforcement Learning
This addresses the scalability and generalization challenges in multi-agent collaborative problems, offering a practical solution for domains like search and rescue and gaming, though it is incremental in leveraging existing structured prediction techniques.
The paper tackles the problem of generalization in cooperative multi-agent reinforcement learning by introducing a structured prediction approach for agent-task assignment, leveraging locality to enable zero-shot generalization to larger problems, significantly outperforming baselines on instances with 5 times more agents and tasks than seen during training.
Effective coordination is crucial to solve multi-agent collaborative (MAC) problems. While centralized reinforcement learning methods can optimally solve small MAC instances, they do not scale to large problems and they fail to generalize to scenarios different from those seen during training. In this paper, we consider MAC problems with some intrinsic notion of locality (e.g., geographic proximity) such that interactions between agents and tasks are locally limited. By leveraging this property, we introduce a novel structured prediction approach to assign agents to tasks. At each step, the assignment is obtained by solving a centralized optimization problem (the inference procedure) whose objective function is parameterized by a learned scoring model. We propose different combinations of inference procedures and scoring models able to represent coordination patterns of increasing complexity. The resulting assignment policy can be efficiently learned on small problem instances and readily reused in problems with more agents and tasks (i.e., zero-shot generalization). We report experimental results on a toy search and rescue problem and on several target selection scenarios in StarCraft: Brood War, in which our model significantly outperforms strong rule-based baselines on instances with 5 times more agents and tasks than those seen during training.