LAUSR

Abstract In a semiconductor manufacturing fabrication, a production line of chemical mechanical polishing (CMP), one of the key semiconductor processes for the wafer planarization, will be arranged by a machine allocation decision twice a day. To prevent machinery malfunctioning in manufacturing, serial scheduling of maintenance operations is generally required over the lifespan of semiconductor processing equipment. However, imprecise information by mostly human decision including condition-life-time type preventive maintenance (PM) and leading operations in CMP may often result in missing the daily target achievement ratio. The issue causes additional leading operation, which is one of the huge expenses caused by a layer change. In this paper, we aim to solve a CMP machine allocation problem. A fuzzy set-based mathematical programming modelling approach is used to synchronize the PM and leading operations. The experimental result shows that the proposed method can sort out the thorny problem, and the target achievement ratio is improved by nearly 5% on average, with short computing time.