Optimizing Cluster Head Selection in Hierarchical Clustered Sensor Network Using Genetic Algorithm

Abstract

Wireless sensor network is the network containing various sensor nodes. Their sole purpose is to aggregate the data from the physical environment and transmit it to the Base Station (BS). This communication will take place wirelessly. Each node has small asset regarding power and bandwidth. Wireless Sensor Network (WSN) applications have expanded for example, ecological detecting, region observing, air contamination checking, machine wellbeing checking, etc. So because of limited battery power, it becomes very important to save the power of these nodes to extend the network lifetime as it is extremely hard to replace batteries. Clustering is a mechanism to extend the lifetime of the WSN. Hierarchical based routing protocols are used to route the traffic from source to the destination via Cluster Head (CH). This sort of routing convention sub-partition the WSN into small clusters and build a hierarchy of nodes. Type-2 fuzzy logic, that follows certain rules, is used to select one node to be the CH so that network life time can be enhanced compare to the simple leach where CH is chosen based on single parameter like remaining energy. Type-2 Fuzzy based system has taken three parameters like density, residual energy and distance for cluster head selection. In proposed work, Genetic based approach has been used. This approach uses selection; crossover and mutation based system so that optimized result can be generated in terms of selection of the most superior sensor node to become a CH. Various performance analysis like dead nodes count, alive nodes count and residual energy have been evaluated. An improvement in case of Genetic based approach as compared to Type-2 Fuzzy based selection criteria and comparison has been observed on different network topologies.