Maximization of WSN Life Using Hybrid Evolutionary Programming

Abstract

Of all the challenges faced by wireless sensor networks (WSN), extending the lifetime of the network has received the most attention from researchers. This issue is critically important, especially when sensors are deployed to areas where it is practically impossible to charge their batteries, which are their only sources of power. Besides the development and deployment of ultra low-power devices, one effective computational approach is to partition the collection of sensors into several disjoint covers, so that each cover includes all targets, and then, activate the sensors of each cover one at a time.. This maximizes the possible disjoint covers with an available number of sensors and can be treated as a set-K cover problem, which has been proven to be NP-complete. Evolutionary programming is a very powerful algorithm that uses mutation as the primary operator for evolution. Hence, mutation defines the quality and time consumed in the final solution computation. We have applied the self adaptive mutation strategy based on hybridization of Gaussian and Cauchy distributions to develop to develop a faster and better solution. One of the limitations associated with the evolutionary process is that it requires definition of the redundancy covers, and therefore, it is difficult to obtain the upper bound of a cover. To solve this problem, a redundancy removal operator that forces the evolution process to find a solution without redundancy is introduced. Through simulations, it is shown that the proposed method maximizes the lifespan of WSNs.