先進生產規劃系統演算法之開發

Abstract

本論文以吳氏所提出之AT（Allowance Time）概念發展一套APS演算法。AT產能規劃機制之概念為以一大於作業純加工時間的時間範圍來決定作業在機台上可以佔用產能的範圍，讓作業在產能分配時，可以在此一範圍內任意佔用機台的產能，因此作業佔用機台產能的位置並沒有固定。故作業在機台上所佔用的產能位置能夠微調，能讓更多的作業分配到機台上加工，達到更高的機台利用率。 吳氏以AT概念提出一套APS演算法，在特定生產情境下，其績效較其他APS軟體為佳。然而該演算法並未對晶圓代工廠中常見的生產特性，包括替代途程（Alternative Routings）、替代機台（Alternative Resources）、批量機台（Batch Resources）等加以探討，因此本研究延續吳氏之成果，將上述三項生產特性考量納入該APS演算法中。 接單管理亦為APS軟體之一項功能，因此本研究亦從APS產能規劃機制之規劃結果，發展一套ATP規劃機制以，幫助業務人員作快速之接單決策。

This research develops an APS (advanced planning and scheduling) algorithm based on the AT concept proposed by Woo. The AT concept advocates that the resource time allocated to a particular job should be larger than its processing time in order that the job can ‘float’ in the capsulated time interval even when it has been planned. That is, the resource time can be reallocated easily so that more jobs can be assigned to the resources and achieve higher utilization of resources. Based on the AT concept, Woo developed an APS algorithm, which has shown better performance than some proprietary APS software in certain production scenarios. Yet, three common properties in IC foundries, alternative routing, alternative resources, and batch resources, are not concerned in Woo’s algorithm. This research enhances Woo’s work by including these three production properties in the proposed APS algorithm. One application of APS software is for the justification of order acceptance. In order to assist sales in quickly making order acceptance decisions, this research also develops an available-to-promise mechanism based on the planning results of APS software.