Double-track particle swarm optimizer for nonlinear constrained optimization problems

Abstract

Particle swarm optimization (PSO) is a popular stochastic algorithm for solving nonlinear optimization problems. It has been used in various industrial applications because of its few parameters and ease of implementation. Various optimal problems must be frequently formed as nonlinear constrained optimization (NCO) problems because of global resource limitations. These problem-specific nonlinear constraints can result in many infeasible regions in the search space. Current PSO-based approaches are inherently designed for unconstrained optimization problems and cannot directly handle the constraint issue; therefore, they add the constraint handling technique (CHT) to account for NCO problems. This study proposes a double-track schema inside PSO called double-track PSO (DTPSO) to address the NCO problem. When an infeasible region blocks a particle, the DTPSO creates a copy of the particle. The original and copied particles utilize different search strategies to move to feasible and infeasible regions. The results indicate that the proposed DTPSO outperforms the seven referenced PSO-based methods in all 33 NCO problems and five recent strong methods in 19 mechanical design problems belonging to the IEEE CEC 2020 competition on real-world single-objective constrained optimization problems. Furthermore, the experiment results demonstrate the quality of the solution and the effective computation of the proposed DTPSO.