2025/05/28 by Jin Xie, Yue Teng, Liang Gao +2 · 1 citation
Engineering · #Scheduling and Optimization Algorithms #Advanced Manufacturing and Logistics Optimization #Assembly Line Balancing Optimization
paper · doi:10.1080/00207543.2025.2508336
Mass customisation is one of the key strategies in the modern business environment, and the flexible job shop is especially suited to mass customisation. Therefore, the flexible job shop scheduling problem (FJSP) has attracted much attention in recent years. Numerous intelligent algorithms have been proposed to solve FJSP, with their performance typically evaluated on benchmarks such as BRdata, BCdata, and DPdata. Despite these efforts, many instances in these benchmarks remain unsolved. In this paper, a hybrid genetic tabu search algorithm (HGTSA) is proposed to address these benchmarks further. First, a code that simultaneously represents machine selection and operation sequence is designed. During the genetic algorithm phase, two specialised crossover operators are developed to guarantee the population diversity; while two mutation operators are devised to prevent premature convergence to local optima. In the tabu search phase, a hybrid neighbourhood structure integrating N8 and k-insertion neighbourhood structures is proposed to enhance the local search capabilities. Comprehensive computational experiments on BRdata, BCdata, and DPdata benchmarks, compared with five state-of-the-art optimisation algorithms, demonstrate the superior performance of HGTSA, both in terms of the quality and stability of the solution. Notably, HGTSA finds a new upper bound for an instance in the DPdata benchmark.