Reinforcement Learning-Enabled Resampling Particle Swarm Optimization for Sensor Relocation in Reconfigurable WSNs
Minghua Wang, Xingbin Wang, Kaiwu Jiang, Bo Fan
Abstract
Aiming to maximize coverage performance and reduce the number of sensors deployed in the reconfigurable wireless sensor networks (RWSNs), in this paper, we first formulate a new cooperative sensing coverage control problem based on the confident information coverage model. Then, inspired by the reinforcement learning and resampling technology, a novel learning automata-based resampling particle swarm optimization (RPSOLA) algorithm is proposed to solve complex multi-peak optimization problem and optimize the cooperative sensing coverage control problem of RWSNs. Experimental results demonstrate that the RPSOLA considerably outperforms other three peer schemes, the RPSO, BASPSO and PSO, in terms of the convergence, coverage rate and sensor redundancy.