Article: Surrogate-based adaptive particle swarm optimisation Journal: International Journal of Wireless and Mobile Computing (IJWMC) 2019 Vol.17 No.2 pp.187 - 195 Abstract: Aiming at the shortcomings of Particle Swarm Optimisation (PSO) in solving complex problems, such as large computation cost and long computation time, the paper proposes a Surrogate-based Adaptive Particle Swarm Optimisation (SAPSO). In the algorithm, PSO carries on the optimisation search through a global exploration population and a local exploitation population. At the same time, a global Gaussian process surrogate and a local one are built based on the history search data to evaluate the two populations approximately. Moreover, some adaptive optimisation strategies have been developed, including Latin Hyper-cube Sampling (LHS) based initial strategy, self-adapting local optimisation, EI sampling-based global optimisation and cooperative searching strategy, which can ensure the balance between the global exploration and the local exploitation validly. The experimental results on benchmark problems show that the proposed algorithm not only can decrease the evaluation costs of the functions validly, but also has good convergence and robustness. Inderscience Publishers - linking academia, business and industry through research

Title: Surrogate-based adaptive particle swarm optimisation

Authors: Lei Zhang; Jing Jie; Hui Zheng; Xiaoli Wu; Shiqing Dai

Addresses: Zhejiang University of Science and Technology, Liuhe Road 318#, Hangzhou City, 310023, Zhejiang Province, China ' Zhejiang University of Science and Technology, Liuhe Road 318#, Hangzhou City, 310023, Zhejiang Province, China ' Zhejiang University of Science and Technology, Liuhe Road 318#, Hangzhou City, 310023, Zhejiang Province, China ' Zhejiang University of Science and Technology, Liuhe Road 318#, Hangzhou City, 310023, Zhejiang Province, China ' Zhejiang University of Science and Technology, Liuhe Road 318#, Hangzhou City, 310023, Zhejiang Province, China

Abstract: Aiming at the shortcomings of Particle Swarm Optimisation (PSO) in solving complex problems, such as large computation cost and long computation time, the paper proposes a Surrogate-based Adaptive Particle Swarm Optimisation (SAPSO). In the algorithm, PSO carries on the optimisation search through a global exploration population and a local exploitation population. At the same time, a global Gaussian process surrogate and a local one are built based on the history search data to evaluate the two populations approximately. Moreover, some adaptive optimisation strategies have been developed, including Latin Hyper-cube Sampling (LHS) based initial strategy, self-adapting local optimisation, EI sampling-based global optimisation and cooperative searching strategy, which can ensure the balance between the global exploration and the local exploitation validly. The experimental results on benchmark problems show that the proposed algorithm not only can decrease the evaluation costs of the functions validly, but also has good convergence and robustness.

Keywords: swarm intelligence; particle swarm optimisation; surrogate; Gaussian process.

DOI: 10.1504/IJWMC.2019.101430

International Journal of Wireless and Mobile Computing, 2019 Vol.17 No.2, pp.187 - 195

Received: 26 Nov 2018
Accepted: 08 Apr 2019
Published online: 07 Aug 2019 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article

Title: Surrogate-based adaptive particle swarm optimisation

Keep up-to-date