The Performance of Genetic Algorithms in Medium Access Control Protocols for Wireless Sensor Networks

Abstract This paper evaluates the performance of the Particle Swarm Optimization (PSO) scheme, a genetic algorithm, as used in Medium Access Control (MAC) protocols for Wireless Sensor Networks (WSNs). In doing so, the protocol is implemented and compared to round-robin, a contention-free protocol that makes sensor nodes take turns in sending their data packet to the sink, and a p-persistent CSMA protocol with a p-value of 0.01. The concept of p-persistence as used in the CSMA protocol is employed to tweak the PSO scheme to form a fourth protocol that is used in the evaluation process. All four protocols are simulated in OMNET++, a network simulator, and performance analysis is done based on the throughput and transmission delays of the various protocols. It was discovered that the PSO scheme performed poorly in terms of throughput in highly populated networks with high traffic. The modified PSO scheme with p-persistence performed better than the original PSO scheme in terms of throughput for densely populated networks with high traffic. On low populated networks with low traffic, the original PSO scheme showed the best throughput performance. Round-robin displayed higher throughput performance than the other three protocols for both low and high traffic networks with a high node population.