| 摘要: | 訂單組合排程問題,即考慮終止、暫停、提前與延遲顧客訂單情形的多筆訂單排程,已經成為今日晶圓廠重要的議題,在晶圓製造業動態和複雜的產業特性下更顯得重要。如何儘可能地快速回應顧客並進行排程並縮短完成時間將成為一個十分迫切的工作。再者,完成訂單的製造活動會消耗資源(包含設備與機台),而資源本身又常是稀少與昂貴並有著資源衝突的情形.以上時間和成本間的取捨問題將會影響排程中的資源配置. 在本研究中,將以一個事件驅動的方法將用來發展一個時間和成本取捨的訂單組合排程與再排程的決策架構,並應用於一晶圓製造廠中.這個決策架構將協助管理同時多筆訂單並解決資源配置的問題.進一步可以將本架構分成監控與控制、時間估計、成本估算與排程衡量四個模組.而這四個模組將以高階裴氏網，作業基礎成本制與TOPSIS三種方法建構之。 本研究的優點包括:1.將多筆訂單與其包含活動之作一正規的描述;2.以圖形呈現訂單之活動與資源受限的情形;3.支援模擬與分析的作業性語法;4.可調整訂單組合或補充、再分配資源的架構 Order portfolio scheduling problem, i.e. multiple orders scheduling problem considering terminate, suspend, preempt, and postpone of customer orders have become a critical issue in today’s semiconductor industry, especially, with the dynamic and complex characteristic. How to respond to customer as soon as possible and schedule orders in a rolling horizon without large makespan (completion time) have been an emergent task. Moreover, the activities to complete orders consume resource (equipment, machine, and labor…etc), and the resource may be scared and expensive. Not only the cost of consume resource is calculated, but also the resource contention occurs. Above time and cost trade off issue influence the resource allocation of production scheduling. In this thesis, an even-driven approach is adopted to develop a tradeoff framework for order portfolio scheduling/rescheduling in a semiconductor fab. This framework is proposed for assisting the management of multiple orders to resolve the allocation of a variety of resources. To provide the solution for the problem, four modules are built into the system: Monitoring and Control, Makespan Estimation, Cost Calculation, and Schedule Evaluation. In order to construct these modules, activity based costing (ABC), High Level Petri Net (HLPN) is employed in tandem. In addition, how to measure and select a preferred feasible schedule will be a dilemma. Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) solves such multi-attribute decision making (MADM) problem. The advantage of this system include:(1) the formal description of multiple orders of which different activities may occur concurrently;(2) the graphical representation of activity of orders with resource constraint, (3) the operational semantics for supporting simulation and analysis, (4) In response to the modification of current order portfolio or supplementary resource being added (removed), reallocating resource to the new set of orders |
