Deciding Near-optimal WSN parameters using Evolutionary Algorithm

We study WSN design using GA technique. The WSN parameters of interest here are communication and operation energy, uniformity in cluster distribution, cluster overlapping, average communication per cluster head, effect of cluster density on communication energy and number of active sensors out of range using our proposed model. The experimental result of this study indicates that the value of radio communication radius has direct effect on cluster overlapping and distribution uniformity of sensors, communication energy, number of active sensors out of range and cluster heads without any member. The optimal value of radio communication radius is not related to the network size and type of deployment. It is related to the density of sensor deployment. In all the network situations we have studied, random deployment has better performance compared to grid deployment.

Keywords:

Wireless sensor network , cluster head , genetic algorithm , active sensor , base station