A genetic algorithm with path-relinking for operation sequencing in CAPP

• Published in: International journal of advanced manufacturing technology Vol. 125; no. 7-8; pp. 3667 - 3690
• Main Authors: Dou, Jianping, Wang, Shuai, Zhang, Canran, Shi, Yunde
• Format: Journal Article
• Language: English
• Published: London Springer London 01.04.2023; Springer Nature B.V
• Subjects: Ant colony optimization; CAE) and Design; Computer-Aided Engineering (CAD; Crossovers; Engineering; Feasibility; Genetic algorithms; Heuristic methods; Industrial and Production Engineering; Mechanical Engineering; Media Management; Mutation; Original Article; Particle swarm optimization; Permutations; Polynomials; Precedence constraints; Operation sequencing; Dynamic programming; Genetic algorithm; Path-relinking; Graph algorithm
• ISSN: 0268-3768; 1433-3015
• DOI: 10.1007/s00170-023-10907-2

Due to the rapid change of customized products in Industry 4.0, the operation sequencing (OS) as a core function of CAPP system is frequently needed. In this paper, a novel path-relinking genetic algorithm (PR-GA) is proposed to find the minimal-cost solution of the OS problem. In the PR-GA, the chromosome records the feasible operation sequence (FOS) satisfying the precedence constraints of operations via a permutation, and the designed crossover and mutation involve chromosomes and ensure their feasibility. For a given FOS, the optimal manufacturing resources for every operation are identified by a polynomial-time graph algorithm including a graph building procedure and a dynamic programing procedure. Thus, the PR-GA focuses on searching promising FOSs, and uses path-relinking as a local search around paths between elitists. Moreover, a new framework of GA is established and is characterized by avoiding tuning crossover and mutation rates of the common GA. The PR-GA is compared with state-of-the-art metaheuristic algorithms (MAs) including ant colony optimization, particle swarm optimization, two recent GAs, and a recent exact method to verify its effectiveness and efficiency. The comparison results show that the PR-GA outperforms existing MAs for solution quality, and illustrates the PR-GA’s promising efficiency and robust global search ability.