本論文旨在探討半導體產業下機台維護保養內重要的一環-工具校準。半導體機台設備的價格昂貴，且折舊快，再加上具有高度工具需求和系統的複雜性，為了使設備提升效能和保持穩定的產能間取得平衡，預防性維護保養(PreventiveMaintenance)在半導體製造中變得不可或缺且具有高度挑戰性。在本文中提出一個整數規劃(Integer programming)的數學模型，並用於維護保養下的工具校準，目的是為了使工具不可用時間最小化，並能順利完成製造廠與設備供應商已規劃好的維護保養。此模型數學模型考量到了許多重要的因素，例如:在時間內工具必須進行一次校準、多種類型的工具，以及工具校準與訂單和維護保養間的關係，並轉化為限制式，藉此更合乎真實案例，以求得最佳解。 This thesis discuss the important part of preventive maintenance in semiconductor industry which is tool calibration.Semiconductor equipment is very expensive and depreciates quickly also possess high tool requirements and system complexity.In order to achieve a balance between improving equipment performance and maintaining stable production capacity,Preventive maintenance became indispensable and highly challenging semiconductor manufacturing.In this thesis, provides a non-linear integer programming mathematical model used for tool maintenance under maintenance.The purpose is to minimize the unavailable time of tools and successfully complete the process between manufacturers and equipment suppliers of maintenance they have been planned.The mathematical model of this thesis considers many important factors,such as tools must be calibrated once in a time, multiple types of tools,and the relationship between tool calibration and orders, then turn intoconstraints of model,which are more close to realistic cases and find the best solution.
