A novel constrained bare-bones particle swarm optimization
Yuanxia Shen, Jian Chen, Chuanhua Zeng, Bin Ji · 2016
Particle swarm optimization (PSO) has been applied to nonlinear constrained problems. However, PSO may easily get trapped in the local optima when solving complex problems and suffers from the setting of learning parameters. In order to improve convergence accuracy of solutions, a hierarchical learning bare-bones PSO is proposed, named HLBPSO, for dealing with constrained optimization problems. HLBPSO adopts a hierarchical learning strategy to maintain population diversity. In HLBPSO, an archive is used to store the accepted infeasible solutions and auxiliary operations are introduced to help accepted infeasible solutions to enter into the feasible region. Experiments were conducted on constrained benchmark problems. The experimental results showed that HLBPSO performs better than four other related works.