An efficient problem-specific evolutionary algorithm for flexible job shop scheduling problem with specific workers in highly customised manufacturing systems

The integrated optimisation problem of production scheduling and workforce scheduling has emerged as a critical challenge in modern manufacturing systems. Although existing research predominantly addresses single-tasking workers capable of handling one operation at a time, the scheduling complexity introduced by multitasking workers performing multiple operations at a time remains understudied. This gap is particularly significant in highly customised industries such as shipbuilding and aerospace manufacturing, where the versatility of the workforce substantially impacts production efficiency. This study investigates the flexible job-shop scheduling problem with multitasking workers (FJSP-MW), proposing a genetic algorithm with knowledge-based local search (GALS). The proposed algorithm incorporates two key innovations: (1) a disjunctive graph model for FJSP-MW with a total weighted tardiness (TWT) objective function and (2) problem-specific neighbourhood structures based on critical paths. Comprehensive experiments evaluate the algorithm's performance using 20 instances and a case study. The results of the case study demonstrate significant improvements; reductions of 32.14% in TWT and 39.02% in makespan are obtained compared to the original scheduling solution. The results confirm that GALS outperforms state-of-the-art algorithms in solution quality and convergence speed.