Overlapping cluster control mechanism for Particle Swarm Optimization-based clustering algorithm

Its believed that Clustering is as a good solution to the needs of energy efficiency as well as scalability on Wireless Sensor Networks (WSN). Many clustering algorithms have been proposed by researchers that generally ask each node join only to one cluster to minimize energy consumption. However, some applications need some nodes to affiliate to more than one cluster. We propose a mechanism that be inserted in existing clustering algorithms so they able to control the overlapping between clusters without reduce the ability to maintain energy efficiency. In the paper, the mechanism is implemented on a Particle Swarm Optimization (PSO)-based clustering algorithm. PSO is a lightweight heuristic optimization method of computing and quickly achieves convergence that very suitable for WSN system that limited in computation resource. Evaluation shows that the proposed mechanism has succeeded adding to PSO-based clustering an ability to control the overlapping among cluster with a very little degradation of performance on maintaining the balance of energy consumption.