Production costs and total completion time minimization for three-stage mixed-model assembly job shop scheduling with lot streaming and batch transfer

Assembly scheduling problems have broad applications in discrete manufacturing enterprises and hence gained extensive attention in academics. However, some important and actual factors including lot streaming, batch transfer of components and mixed-model assembly are yet not well studied. To adapt to actual production and enhance competitiveness, a three-stage mixed-model assembly job shop scheduling with lot streaming and batch transfer is investigated in this work. To tackle this problem, spatial-temporal links among three stages with differentiated production lot size are established and a novel mathematical model is formulated to minimize total completion time and production costs including setup costs, transport costs, and inventory costs. Besides, an improved multi-objective coevolutionary simulated annealing algorithm is developed. To promote the coevolution of processing, transfer, and assembly sub-problems, the multi-attribute combination rule, similarity rule, and consistency rule are respectively designed for each sub-problem. An adaptive execution strategy of sub-problems is trained by Q-learning to maximize the overall benefits. Experimental results show that both the discovered rule and the adaptive strategy are effective, and the proposed algorithm significantly outperforms seven competitive multi-objective algorithms in fixing the studied problem. More importantly, the developed production mode of "processing with lot streaming-batch transfer-mixed-model assembly considering pull production" has a better comprehensive performance than the other three actual production modes.