New fault tolerant strategy of wireless sensor network

Abstract

The wireless sensor network is enormously used in a random way in different environments to execute diverse applications and tasks. Because of its sensitivity, various research projects have been conducted with the aim of finding solutions in the presence of failures. New versions of K-means And Traveling Salesman Problem based mobility protocol, based on K-means clustering and the approximate solution for Traveling Salesman Problem by using the simple local search algorithm “2-Opt”, have been proposed in order to obtain real time solutions once the problem is disturbed by the breakdowns of some nodes. In this paper, we aim to use a new strategy by implementing algorithms employed for probabilistic combinatorial optimization problems resolution, such as exact methods. We propose then a new strategy that navigates the mobile sink to go through the cluster centers according to the optimized path by the implementation of an exact resolution of this problem using the “branch and bound” technique. Simulation results have demonstrated that although this exact solution is very expensive, it can be so interesting when the risk of sensor breakdowns becomes considerable and the number of cluster centers will be reduced, and that it outperforms the original solution of K-means And Traveling Salesman Problem based mobility protocol in terms of quality despite the increase in complexity.