Node Deployment Optimization for Wireless Sensor Networks Based on Virtual Force-Directed Particle Swarm Optimization Algorithm and Evidence Theory
Liangshun Wu, Junsuo Qu, Haonan Shi, Pengfei Li
Abstract
Wireless sensor network deployment should be optimized to maximize network coverage. The D-S evidence theory is an effective means of information fusion that can handle not only uncertainty and inconsistency, but also ambiguity and instability. This work develops a node sensing probability model based on D-S evidence. When there are major evidence disputes, the priority factor is introduced to reassign the sensing probability, with the purpose of addressing the issue of the traditional D-S evidence theory aggregation rule not conforming to the actual scenario and producing an erroneous result. For optimizing node deployment, a virtual force-directed particle swarm optimization approach is proposed, and the optimization goal is to maximize network coverage. The approach employs the virtual force algorithm, whose virtual forces are fine-tuned by the sensing probability. The sensing probability is fused by D-S evidence to drive particle swarm evolution and accelerate convergence. The simulation results show that the virtual force-directed particle swarm optimization approach improves network coverage while taking less time.