Method of job shop scheduling considering reworking and reprocessing based on proactive approach

Abstract

This paper is concerned with performing job shop scheduling considering discrete uncertainty based on proactive approach. In production of large products, there often are cases where a work-in-process needs to undergo a reworking or a reprocessing depending on the result of an inspection process. Generally speaking, it is natural to cope with this type of uncertainty based on reactive approach, since it is discrete and makes a drastic change on the original production scenario. However, this may result in an unsatisfactory schedule which requires unreasonable overtime works or fails to keep due dates, etc., because the initial schedule is generated without considering the uncertainty and is updated considering only the information that is certain at the time of updating it. For this reason, we developed the following method based on proactive approach: (i) a set of consistent schedules each of which corresponds to a production scenario determined by an inspection result is generated by using genetic algorithm so that the weighted sum of overtime work, tardiness and makespan or earliness is minimized; (ii) the production starts based on the initial schedule that corresponds to the scenario in which no reworkings/reprocessings are necessary, and then the schedule is switched by the one which corresponds to the new scenario each time when it turns out that a reworking/reprocessing needs to be carried out. Comparison to reactive scheduling approach using numerical examples showed effectiveness of the proposed method.