Dynamic overlapping clustering for wireless sensor networks based-on particle swarm optimization

In the recent years, Wireless Sensor Network (WSN) has been one of the most interesting research topics because of its flexibility and many potential applications. However, in the applications, there are still resources constraints, including: energy, computation, and bandwidth. It is believed that clustering is the best solution for the need of energy efficiency and scalability. In order to reach the high level of energy efficiencies, mostly, the clustering algorithms avoid the possibility of overlap between clusters. But in fact, there are several applications that need the occurrence of overlaps between clusters. In this paper, we propose a Particle Swarm Optimization (PSO)-based Clustering algorithm that has capability to control the overlap between clusters but still it has an ability to reach energy efficiency. PSO is chosen because it has a light computation and can quickly reach convergence. This proposed algorithm performance is analytically and experimentally compared with clustering on LEACH. The result of the test shows that this proposed algorithm has a capability to control the rate of overlapping degree linearly. The testing on the PSO for clustering also shows the better performance than on LEACH, although there are a few problems related to its complexity.