| Abstract: | 限制驅導排程是限制理論運用於現場排程中的技術，其主要精神為充份利用瓶頸或限制資源，以確保整體產能輸出能夠最大化，為了使限制資源產能被充分發揮，其主要技巧在於緩衝的管理。本研究以零工式平行機台排程型態為基礎，提出使用基因演算法配合和限制驅導排程的技術，以質性與量性瓶頸資源為優先，對於企業多目標的營運策略，求解出一套完整的排程規劃。現實企業追求營運目標時，不單單是只有單目標，而是朝著多目標去追尋，但常會受到不同型態的瓶頸阻擾，使的規劃變的困難，大多目標只針對量性目標如：製距、交期等，但沒考慮到的質性目標如：訂單潛在利潤、以往的交易歷史等，這些目標也是在現實世界中也是很重要的。本研究主要於在零工式平行機台的架構下，發展一套限制驅導排程法，並利用基因演算法針對有限的限制資源做最大化利用，針對以往對於緩衝長度多半是以經驗值去做判斷去加以做修改，並保護瓶頸作業，績效考量下，考量企業多目標的策略包含質性與量性策略，並以排程規劃人員設計出一套合適的相關權重，在諸多考慮下，才能建構出本研究的多目標適應函數。經由本研究證明，使用限制驅導排程結合基因演算法，會比一般排程方法的排程解更為明確。
Drum-Buffer-Rope (DBR) is Theory of Constraints (TOC) use scheduling theory of technology to the field, main spirit for take advantage of bottlenecks or constraint resources, in order to ensure that the overall capacity to maximize output, the main trick is to buffer management. This study is based on job-shop parallel-machine scheduling, proposed use Genetic Algorithms(GA) combine with Drum-Buffer-Rope, as a priority in the quantitative bottlemeck and qualitative bottleneck, for multi-enterprise business strategy goals, solving a set of complete schedule planning.When the reality of companies pursuing business goals, not just only a single goal, but moving more goals to pursue it, but usually obstruction with different types of bottlencks, it make scheduling become very difficult, most goals for the quantitative bottlemeck, for example: manufacturing length and due date, but not consider the qualitative bottleneck, for example:orders potential profits and previous transaction history, this goals also very important in the word.This study focuses on job-shop parallel-machine scheduling, to develop a DBR method, and take advantage of GA to maximize the use of resources for a finite constraints, making changes to the buffer length for most of the past experience that do it to determine and protection of bottleneck, consideration of multi-objective strategy includes quantitative strategy and qualitative strategy and schedule planners to design an appropriate relevant weights, also construct the multi adaptation function with this study in many consideration. The results will be identified to the GA combine DBR better than general scheduling method. |
