Scheduling optimization for laminated door machining shop based on improved genetic algorithm

Abstract

In the digital transformation of the wooden-door manufacturing industry, material preparation planning and production scheduling directly influence the stability and effectiveness of the manufacturing system. Constructive problem-specific algorithms have been instrumental in solving real-world laminated door machining shop scheduling problem (LDMSSP). LDMSSP is a complex problem that combines a distributed permutation flow-shop scheduling problem and distributed hybrid flow-shop scheduling problem. An improved genetic algorithm fused with the strategies of the improved heuristic algorithm, the local search, variable neighborhood search with multiple critical paths, and the iterated greedy search (IGGA) was proposed for application in the material preparation planning and scheduling optimization to minimize the makespan. Comprehensive design of experiments and statistical analyses were conducted to determine appropriate algorithm parameters and verify the substantial improvement of the IGGA. Experiments conducted on various benchmark instances indicated that IGGA outperformed other metaheuristics in both the best relative deviation index and the average relative deviation index. In the end, the minimal makespan for a real-world case involving the production of 74 laminated doors was 1.1 h with a 17.91% reduction, which further demonstrated the effectiveness of the proposed model and algorithm in solving LDMSSP. It also provided a valuable reference for the rational arrangement of material preparation planning and machining scheduling sequences.