Abstract: The scheduling in a multi-level assembly job shop is a two-stage scheduling problem involving both processing and assembly. The assembly products have different tree-type structures and the assembly operation at each level requires the completion of directed components. Lot-splitting scheduling can improve the production liquidity in a job shop. Thus, it is usually applied in fields such as job shop scheduling. The employment of lot splitting in an assembly job shop requires the coordination of related components and the progress of their sublots. To this end, this paper establishes a lot-splitting scheduling model with the optimization objective being the minimization of the delay cost and the inventory cost in a multi-level assembly job shop. With the consideration of efficiency, this paper establishes a hybrid solution algorithm based on genetic algorithm and priority dispatching rules to solve the two subproblems of lot splitting and lot scheduling. Finally, simulation experiments are conducted to verify the effectiveness of the proposed lot-splitting scheduling algorithm, and the adaptability of the hybrid algorithm to differentiated product structures is evaluated under eight priority dispatching rules and three types of lot-splitting strategies. Experiment results show that the equal size sublot strategy can effectively improve the scheduling performance of the hybrid algorithm under given conditions.

Key words: assembly job shop, tree structured products, lot-splitting scheduling, genetic algorithm, priority dispatching rules