Multi agent model improves team planning in cooperative tasks
MA-WAM: Multi-Agent World-Action Model for Test-Time Planning
Artificial Intelligence
Summary
Cooperative tasks with multiple agents need everyone to act together at the same time, but each agent's action affects the others, making planning hard. The authors created MA-WAM, a system that predicts what will happen when all agents act jointly by understanding their interactions. This method helps choose the best group actions, improving team performance in many simulated tasks. It works quickly without slowing down the action planning.
What this means in practice
- •For robotics teams: Improve coordinated control and planning of multiple robots by better predicting outcomes of joint actions in shared environments.
- •For autonomous vehicle fleets: Enhance decision-making for fleets of self-driving vehicles working together by evaluating the combined effects of their simultaneous maneuvers.
Authors
Guowei Zou, Haitao Wang, Guoxin Wang, Beiwen Zhang, Zhiquan Chen, Guojie Wang, Hejun Wu
Abstract
Multi-agent cooperative tasks require different agents to execute a joint action simultaneously, and each agent's action affects both the observations and responses of the other agents. Hence, a world model is needed to predict the team return resulting from the joint actions of all agents. A naive extension directly applies a single-agent world model to each agent's action when predicting the team return step by step. However, such an extension fails to capture the dependencies among the simultaneous actions of multiple agents. We propose Multi-Agent World-Action Model (MA-WAM), a test-time planning framework that enables a frozen multi-agent flow policy to evaluate futures of candidate joint actions. To our knowledge, MA-WAM is the first test-time world-model planner for multi-agent flow policies. MA-WAM predicts the consequences of each joint action according to cross-agent dependencies and enables efficient candidate scoring. Across 30 offline multi-agent reinforcement learning (MARL) settings on MAMuJoCo, SMAC, and MPE, MA-WAM achieves mean relative gains of 22.0% over direct execution and 25.6% over uniform action selection. Under the standard evaluation protocol on an A100 GPU, MA-WAM adds 12.1 ms, accounting for 2.5% of the measured generation-and-scoring time.