Distributed Cooperative Multi-Agent Reinforcement Learning with Directed Coordination Graph
This work addresses scalability issues in multi-agent systems for domains like resource allocation, but it is incremental as it builds on existing zeroth-order optimization methods.
The paper tackles the problem of high communication costs and poor scalability in distributed cooperative multi-agent reinforcement learning by proposing a new algorithm that uses directed coordination graphs and local value functions without consensus algorithms, showing high scalability in a resource allocation example.
Existing distributed cooperative multi-agent reinforcement learning (MARL) frameworks usually assume undirected coordination graphs and communication graphs while estimating a global reward via consensus algorithms for policy evaluation. Such a framework may induce expensive communication costs and exhibit poor scalability due to requirement of global consensus. In this work, we study MARLs with directed coordination graphs, and propose a distributed RL algorithm where the local policy evaluations are based on local value functions. The local value function of each agent is obtained by local communication with its neighbors through a directed learning-induced communication graph, without using any consensus algorithm. A zeroth-order optimization (ZOO) approach based on parameter perturbation is employed to achieve gradient estimation. By comparing with existing ZOO-based RL algorithms, we show that our proposed distributed RL algorithm guarantees high scalability. A distributed resource allocation example is shown to illustrate the effectiveness of our algorithm.