Sublot-level coordinated scheduling for multi-shop manufacturing system under kitting constraints: a full active schedule driven solution framework

With the increasing prevalence of make-to-order production, manufacturing systems are shifting from isolated single-shop to multiple interdependent shops. In the case of lot streaming, downstream assembly is highly sensitive to kitting completeness, inducing strong sublot-level coupling between fabrication and assembly. Existing studies offer limited support for sublot-level fabrication–assembly coordination and seldom exploit feasible time windows for kitting synchronisation. Motivated by metal tube-frame furniture production, this study investigates a sublot-level coordinated fabrication–assembly scheduling problem (SL-FASP) to minimise total tardiness. The system consists of a hybrid flowshop for component fabrication and a flowshop for assembly. Component lots can be split into intermingled sublots, while assembly is performed at the product-sublot level under kitting constraints. A mixed integer linear programming (MILP) model is formulated, and a full active schedule (FAS)-driven solution framework (FAS_SF) is developed. It constructs kitting-consistent schedules by exploiting earliest–latest feasible time windows and improves them through a co-adaptive iterated greedy (CIG)-based improvement phase with FAS-based reverse-encoding and deterministic re-decoding. Results show that FAS_SF achieves the lowest mean ARPI in most real-world-scale and large-scale configurations compared with state-of-the-art heuristics and reduces average total tardiness and runtime by 60.6% and 98.4%, respectively, compared with CPLEX on small-scale instances.