2010-08-012024-05-18https://scholars.lib.ntu.edu.tw/handle/123456789/699483摘要：動態資源配置與動態預防保養排程能有效因應需求不確定與系統可靠度問題，但現有方法 受限運算複雜性（computational complexity），未能應用於大型系統內。針對現有理論與方法的 缺點，本計畫中的動態決策將考慮：1.即時的系統可靠度狀態與等候線長度；2.系統的未來狀 態變化機率（如：訂單到達時間的機率分配及各資源的損壞、維修發生機率等）。不同於更新 時間間隔較長的現有方法，本計畫將決策時間點（decision epoch）細分至所有系統內動態事件 （如服務完成、訂單到達與設備損壞等）的發生時刻，所以能夠完整考慮其他方法所無法描述 的系統動態特性。 本研究以無限規劃時間隨機動態規劃（ infinite planning horizon stochastic dynamic programming）為方法，進行動態資源配置與預防保養排程最佳化。第一年度中將證明最佳預 防保養與動態資源配置解的單調結構特性（monotone structure），並利用最佳解結構特性實際 撰寫單階段與兩階段系統整合求解軟體；第二年度中將依據上述結果開發大型系統之動態資源 配置與預防保養排程演算方法及軟體工具，第三年將整合大型系統內之預防保養排程及動態資 源配置方法，並發展決策支援系統以提高系統平均產出率並降低庫存與等候成本，且本研究將 以TFT-LCD 製程為例，模擬驗證動態決策方法的強健性（robustness）。除學術貢獻外，也期 望成為產業界進行資源配置與預防保養時的參考。<br> Abstract: Systems with dynamic resource allocation and dynamic preventive maintenance are robust to reliability and demand uncertainty problems. However, due to the computational complexity, it remains unclear how dynamic resource allocation and preventive maintenance should be done in large systems. In this research, we propose a new dynamic method for resource allocation and preventive maintenance scheduling problems in large systems. This new method explicitly considers all dynamic events in the system, including demand uncertainties, queue length distribution, and real time resource reliability status. The decision epoch of this method is the time of every dynamic event in system. Due to the large number of decision epochs, infinite horizon stochastic dynamic programming will be used to solve the resource allocation and preventive maintenance problem. First, we will show the monotone structure of optimal resource allocation and preventive maintenance policies. This monotone structure can facilitate the search of optimal policies in large systems. A new decision support system will then be developed for systems under demand and reliability uncertainties. To verify the robustness of proposed method, TFT-LCD manufacturing systems will be studied using discrete event simulation methods. The goal of this research is to provide a guideline for solving resource allocation and preventive maintenance scheduling problems in industry.隨機動態規劃可靠度需求不確定動態資源配置動態預防保養規劃stochastic dynamic programmingreliabilitydemand uncertaintydynamic resource allocationdynamic preventive maintenance scheduling