MULTI-OBJECTIVE DYNAMIC JOB SHOP SCHEDULING: A SURVEY AND PROSPECTS

Job shop scheduling is a classical combinational optimization problem. Most real-world environments are dynamic and a variety of disruptions may occur unexpectedly, e.g., arrivals of new jobs, machine breakdowns. This paper analyzes the limitations of the static approaches when rescheduling in dynamically changing environments, and the instability problem induced by rescheduling when only based on shop efficiency. Then issues that have appeared in recent years on dynamic job shop scheduling with multiple objectives are described. We categorize the disruptions as data uncertainties and real-time events, and classify the typical objectives in dynamic scheduling into three categories, which are the shop efficiency, schedule robustness and system stability. A new mathematical model of the multi-objective dynamic job shop scheduling problem (MODJSSP) is constructed based on a robust-reactive scheduling approach. A literature review of the state-of-the-art multi-objective dynamic job shop scheduling approaches is provided, and the strength and weakness of different approaches are discussed, and the gaps in current research work are summarized. Experimental studies validate the effectiveness and efficiency of our new mathematical model for MODJSSP.