Dynamic opportunistic maintenance grouping in a lot streaming based job-shop scheduling problem

Abstract

This paper aims to model maintenance planning with a dynamic opportunistic approach for a job-shop production system. One issue in such a system is the positive or negative economic dependency. That is grouping maintenance activities may decrease or increase system costs. Furthermore, many maintenance models consider the planning of maintenance only based on a long-term horizon. While short-term and real circumstances such as system characteristics and constraints, workload, number of available maintenance teams, and variable maintenance cost and time are almost ignored. To address these issues, a rolling-horizon approach based on a long-term maintenance plan is proposed so that subsequent scheduling of maintenance and production activities are performed as events unfold through the time. Hence, we have developed a mixed-integer nonlinear mathematical model to simultaneously make decisions on maintenance selection, maintenance grouping, lot sizing and production scheduling. The objective function includes the costs of preventive and corrective maintenance activities as well as various production costs such as production and setup, tardiness penalty, and safety stock penalty. A self-adaptive Cuckoo Optimization Algorithm has been used to solve the proposed model. Numerical experiments were conducted to demonstrate the validity of the model and investigate the efficiency and effectiveness of the optimization algorithm.