Research on a Wireless Mesh Network Routing Algorithm based on Reinforcement Learning

• 1、School of Electronic Engineering, Beijing University of Posts and Telecommunications, Beijing 100876

Abstract：Aiming at the network routing problem that it is hard to quickly adapt to real-time changing network status in traditional Mesh network, which is easy to cause the uneven and inadequate network resource allocation and degradation of network performance, this paper proposes an intelligent wireless Mesh network routing protocol based on reinforcement learning. This paper analyzes the existing use of reinforcement learning methods to assist in solving the routing problems in wireless networks. Most of the related researches did not fully consider the lack of load balancing routing performance of data packets, including interfering node loads, and use reinforcement learning technology to design in line with wireless Mesh The intelligent routing algorithm of the network routing feature has designed a reward function including link interference and node load, so that it can continuously learn according to the changes of the network status to adapt to the dynamic changes of the wireless medium and network topology, and dynamically adjust the routing strategy to Select the next hop relay node to forward the data packet, and finally achieve the effect of effectively avoiding the heavy load area of the network to achieve load balancing. After simulation verification, the routing algorithm proposed in this paper effectively improves the network performance such as network throughput, delay and packet loss rate.

Keywords： Computer network Wireless mesh networks Routing Machine learning Reinforcement learning Load balancing Regional load aware