Optimizing makespan and stability risks in job shop scheduling

Abstract

In real-world manufacturing environments, the execution of a schedule often encounters uncertain events, which will bring the risks of performance deterioration and production system instability. This study addresses the optimization of risks both in performance and stability for the job shop scheduling under random machine breakdowns, in which three objectives: makespan, makespan risk and stability risk are considered at the same time. The buffering approach under the limited predictive makespan will be proposed and used to generate predictive schedules, which allows inserting additional idle time to control the risks. By utilizing the available information about the relationship between the risks and the random machine breakdowns, we have developed two kinds of operation-block based buffering strategies. In order to meet the decision makers with different risk preferences, a multi-objective predictive scheduling algorithm with the proposed buffering strategies is developed to generate a Pareto solution set. Extensive experimental results indicate that, compared with the existing methods, the proposed method can provide a better Pareto solution set in terms of both the diversity and the convergence.