An infeasible domain memetic algorithm for the assembly job shop scheduling with part sharing

This paper investigates the assembly job shop scheduling problem with part sharing (AJSP-PS), with the objective of minimising the makespan by coordinating processing, assembly, and shared-component allocation decisions. To capture the interchangeability induced by part sharing, a processing operation network and a process-constraint swap (PCS) operator are developed, based on which a reinforced PCS–MPJM neighbourhood is constructed. An infeasible-domain memetic algorithm (IFDMA) is then proposed, integrating minimum-move-distance repair, active decoding with shared-part allocation, adaptive tabu search, and diversity-aware evolutionary mechanisms. Computational experiments on 15 benchmark instances show that IFDMA outperforms the compared algorithms in solution quality and robustness.