Title: Different force laws driving artificial physics optimisation algorithm for constrained optimisation problem

Authors: Liping Xie; Jianchao Zeng; Jian Yin

Addresses: Division of Industrial and System Engineering and Complex System and Computational Intelligence Laboratory, Taiyuan University of Science and Technology, No. 66 Waliu Road, Wanbailin District, Taiyuan, Shanxi 030024, China ' Division of Industrial and System Engineering and Complex System and Computational Intelligence Laboratory, Taiyuan University of Science and Technology, No. 66 Waliu Road, Wanbailin District, Taiyuan, Shanxi 030024, China ' Institute of Computer science and technology, Taiyuan University of Science and Technology, No. 66 Waliu Road, Wanbailin District, Taiyuan, Shanxi 030024, China

Abstract: Inspired by physical force, Artificial Physics Optimisation (APO) algorithm is a novel stochastic based on a physicomimetics framework. Driven by virtual force, a population of sample individuals searches for a global optimum in the problem space. The force law is a key problem associated with the performance of APO algorithm significantly. In the paper, an APO algorithm with the Feasibility and Dominance (FAD) method (FAD-APO) is employed to solve constrained optimisation problems. Three different force laws are constructed between the feasible individuals and infeasible individuals, which drive all individuals to search for a global optimum in the constrained problem space. Simulation results show that FAD3-APO algorithm may generally perform better than FAD1-APO and FAD2-APO; it is the most stable and effective among the three versions of FAD-APO algorithms. Meanwhile, a comparison with other population-based heuristics shows that the FAD-APO algorithm is competitive on some test function.

Keywords: artificial physics optimisation; APO; constrained optimisation; force law; feasibility-based rules; physicomimetics; virtual force.

DOI: 10.1504/IJWMC.2015.073099

International Journal of Wireless and Mobile Computing, 2015 Vol.9 No.3, pp.290 - 299

Received: 15 Apr 2015
Accepted: 15 May 2015
Published online: 19 Nov 2015 *

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