Partially Overlapped Channel Assignment Using Discrete Particle Swarm Optimization

Abstract

Wireless mesh networks are multi-hop wireless networks whose capacity decreases very fast due to interference from parallel transmissions on the same channel when using orthogonal channels. In order to further eliminate network interference and improve network capacity, partially overlapped channels are used to perform channel assignment, and a discrete particle swarm optimization (DPSO) based channel assignment algorithm using partially overlapped channels is proposed in this paper. Channel assignment for all links is mapped to position of particle, and DPSO is used to evolve and produce better channel assignment solutions. Simulation results show that network performances can be dramatically improved by properly utilizing partially overlapped channels, for example, the average end-to-end delay and the average packet loss ratio both can be decreased by at least 15 percent, and the network throughput can be improved by 12 percent or more.