Online single batch machine scheduling with linear setup times and incompatible jobs for autoclave molding manufacturing

This paper investigates an online single batch machine scheduling problem for autoclave molding in composite materials manufacturing, in which the batches of jobs to be performed are modified by the new arrival jobs. The single batch machine and jobs are represented by rectangles, and jobs can be processed as a batch if the two-dimensional constraints are fulfilled, i.e., any pair of two jobs in a batch cannot overlap each other. Preemption is not allowed. Besides, the linear setup time and family incompatibility constraints are considered. The goal is to minimize the maximum completion time of batches, the makespan. To tackle this problem, two online approaches are proposed. The first one called Greedy Exact (GE) is to reschedule all unprocessed jobs as well as the new arrival jobs by a mixed-integer programming (MIP) upon a decision point. The second is a tailored rule-based algorithm (RBA), which is devised to balance the trade-off between the waste of machine resources and machine space. Both methods are compared to an offline exact model (EM), which assumes that jobs’ release times are known before starting the production. Experimental simulations are conducted to verify the effectiveness and efficiency of proposed approaches and analyze the impact of instances’ features on schedules. Besides, the management implications are provided for decision-makers to achieve managerial implications and insights.