Co-learning Multi-agent Congestion Control for High-Speed Networks

In this paper, we present an adaptive multi-agent reinforcement learning method for solving congestion control problems on high-speed networks. Traditional reactive congestion control regulates source rate in terms of queue length restricted to a predefined threshold. However, the determination of the congested threshold and sending rate is difficult and inaccurate due to the dynamic nature of the networks. We proposed a simple and robust Co-learning Multi-agent Congestion Controller (CMCC), which consists of two sub-systems: a long-term policy evaluator and a short-term rate selector incorporated with a co-learning reinforcement signal to solve the problem. By means of learning, CMCC can always take correct actions adaptively to regulate source flow under time-varying environments. Simulation results showed the proposed approach can promote the system utilization and decrease packet losses simultaneously. Key-Words: Congestion Control, Reinforcement Learning, Co-learning.