A new hybrid GA−ACO−PSO algorithm for solving various engineering design problems

• Published in: International journal of computer mathematics Vol. 96; no. 5; pp. 883 - 919
• Main Authors: Tam, Jun Hui, Ong, Zhi Chao, Ismail, Zubaidah, Ang, Bee Chin, Khoo, Shin Yee
• Format: Journal Article
• Language: English
• Published: Abingdon Taylor & Francis 04.05.2019; Taylor & Francis Ltd
• Subjects: 65Kxx mathematical programming; 80M50 optimization; 90C26 non-convex programming; 90C27 combinatorial optimization; Algorithms; Ant colony optimization; constrained engineering problem; Crossovers; Design engineering; Entrapment; genetic algorithm; Genetic algorithms; global optimization; hybrid algorithm; Linearity; Mutation; Operators (mathematics); optimization and variational techniques; Particle swarm optimization; Searching
• ISSN: 0020-7160; 1029-0265
• DOI: 10.1080/00207160.2018.1463438

The intention of this hybridization is to further enhance the exploratory and exploitative search capabilities involving simple concepts. The proposed algorithm adopts the combined discrete and continuous probability distribution scheme of ant colony optimization (ACO) to specifically assist genetic algorithm in the aspect of exploratory search. Besides, distinctive crossover and mutation operators are introduced, in which, two types of mutation operators, namely, standard mutation and refined mutation are suggested. In early iterations, standard mutation is utilized collaboratively with the concept of unrepeated tours of ACO to evade local entrapment, while refined mutation is used in later iterations to supplement the exploitative search, which is mainly controlled by particle swarm optimization. The proposed method has been validated in solving test functions and well-known engineering design problems. It exhibits a great global search capability even in the presence of non-linearity, multimodality and constraints, involving a large number of dimensions as well as large search areas.