Safe Model-Based Multi-Agent Mean-Field Reinforcement Learning
This work addresses scalability and safety challenges in multi-agent systems for applications like shared mobility, offering a novel approach to handle constraints in large populations.
The paper tackles the problem of coordinating many agents under global constraints, such as capacity or coverage requirements, by proposing Safe-M^3-UCRL, a model-based mean-field reinforcement learning algorithm that ensures safe policies with unknown transitions, and demonstrates its effectiveness in simulations on a vehicle repositioning task using real trajectory data from Shenzhen.
Many applications, e.g., in shared mobility, require coordinating a large number of agents. Mean-field reinforcement learning addresses the resulting scalability challenge by optimizing the policy of a representative agent interacting with the infinite population of identical agents instead of considering individual pairwise interactions. In this paper, we address an important generalization where there exist global constraints on the distribution of agents (e.g., requiring capacity constraints or minimum coverage requirements to be met). We propose Safe-M$^3$-UCRL, the first model-based mean-field reinforcement learning algorithm that attains safe policies even in the case of unknown transitions. As a key ingredient, it uses epistemic uncertainty in the transition model within a log-barrier approach to ensure pessimistic constraints satisfaction with high probability. Beyond the synthetic swarm motion benchmark, we showcase Safe-M$^3$-UCRL on the vehicle repositioning problem faced by many shared mobility operators and evaluate its performance through simulations built on vehicle trajectory data from a service provider in Shenzhen. Our algorithm effectively meets the demand in critical areas while ensuring service accessibility in regions with low demand.