An adaptive switchover hybrid particle swarm optimization algorithm with local search strategy for constrained optimization problems

• DOI: 10.1016/j.engappai.2020.103771
• Publication Date: 2020-08-01T16:20:32Z
• Authors: Zhao Liu; Zhiwei Qin; Ping Zhu; Han Li
• Keywords: Derivative-Free Optimization; Premature convergence; Benchmark (surveying); Engineering optimization; Differential Evolution

Abstract Practical engineering optimization problems are almost constrained optimization problems and difficult to be solved effectively, therefore, how to handle these problems has attracted more and more attention. Particle swarm optimization (PSO) is one of the most popular algorithms in solving the complicated optimization problems due to its relatively strong global optimization capability and low requirement for computing resources. However, PSO is easy to converge prematurely like other swarm intelligence algorithms due to the loss of diversity among particles. This article proposes an adaptive switchover hybrid PSO framework with local search process (ASHPSO), which adaptively switches the optimization searching process between the standard PSO and the differential evolution (DE) modified by a full dimension crossover strategy to avoid the premature convergence problem. Moreover, a local search strategy is employed to improve the boundary search capability of PSO in consideration of the engineering problems characteristics. Experiments on 28 well-known benchmark functions, 5 engineering problems and a full vehicle multi-disciplinary optimization problem demonstrate the effectiveness of the proposed algorithm compared with other hybrid variants.