Deployment optimization for 3D industrial wireless sensor networks based on particle swarm optimizers with distributed parallelism

For wireless sensor networks (WSNs), traditional studies on deployment problems center upon 2D plane or 3D full space. However, practical situations are more complex, and simplifications may hinder real-world application. In this paper, we study the scenario of a 3D industrial space with obstacles (i.e., devices). Heterogeneous directional sensor nodes and relay nodes are deployed to maximize coverage and prolong lifetime, respectively. Specifically, sensor nodes are deployed for the maximization of coverage; after the positions of sensor nodes are generated, we deploy relay nodes to maximize the lifetime. A modified 3D coverage model and a lifetime model with reliability constraint are presented to facilitate the mathematical analysis of the deployment problem. For the NP-hard deployment problem, two particle swarm optimizers, the cooperative coevolutionary particle swarm optimization 2 (CCPSO2) and the comprehensive learning particle swarm optimizer (CLPSO), are employed. To reduce the computation time, distributed parallelism based on message passing interface (MPI) is conducted by dividing the 3D deployment space. Extensive experimentations are conducted by using various numbers of sensor nodes and relay nodes, and thorough understandings are obtained w.r.t. both the deployment problem and the optimizers.