Modelling and scheduling preventative maintenance in semiconductor manufacturing industry with MAs

Abstract

Preventive Maintenance (PM) scheduling is a valuable tool to improve reliability and productivity in semiconductor manufacturing. In this paper, we propose an expanded PM Scheduling (PMS) model, which takes into consideration both tools' technical factors and production information. This model aims to schedule PM tasks given the criteria of maximum throughputs, minimum Work-In-Process (WIP) and minimum lateness of PM tasks under operation and resource constraints. In order to precisely model the joint impact of PM tasks on the throughput of cluster tools, we develop an innovative method to formulate availability function of cluster tools. Then, a Memetic Algorithm (MA) is developed to find a best solution for the PMS problem. Consequently, we can treat PMS problem with fine granularity of planning time unit and long scheduling horizon for large systems composed of complex integrated cluster tools in semiconductor manufacturing systems. Practical case study has demonstrated the performance of the proposed method.