Improved discrete particle swarm optimization for solving the practical sensors deployment

Abstract

Sensors deployment has played an important role in many engineering applications, and the key goal is aimed at achieving an optimal surveillance region with a set of sensors. In this paper, a probabilistic strategy was chosen as the sensing model and a Gaussian probability distribution was employed, furthermore an accumulative probability for all the utilized sensors was presented and an optimal deployment on meshed planar grid was proposed. It was proved that the deployment problem was NP-complete, and an approach for approximating this solution should be resorted to intelligent methods. Particle swarm optimization (PSO) was a widely used artificial intelligent tool, and hereby an improved discrete PSO (DPSO) was proposed for solving the deployment problem, and which was based on integer coding, and the initialization, positions and velocities updating were distinct with the traditional PSO. In final, the deployment was investigated respectively by using uniform sensors (binary coding problem) and combinational sensors (multivariate integer coding problem), which were indicated to the core structure of proposed DPSO.

http://ift.tt/2pulJBd