物料規劃與配置為供應鏈規劃主要功能之一,目的在於各種規劃條件的限制之下,如何將各種可行的供給配置到各個獨立的需求,達成企業的需求目標。TFT-LCD傳統物料配置的做法,主要將個別的供給與需求進行優先權排序,進而進行配置的工作,但此舉往往只能求得有效解,而不是最佳解,不能完全達成企業的需求目標,降低更多的製造成本。 由於,基因演算法求解問題時,需花費大量的計算及時間的成本,當問題領域愈大,求解時間也將成指數成長,有鑑於此,本研究使用限制式基因演算,結合推理機制向前檢查 (Forward Checking)的概念,應用在TFT-LCD產業,解決物料規劃與配置最佳化的問題,求取產量最大化。
Material planning and allocation are main parts of supply chain management. It aims to locate various supplies to individual demand to satisfy enterprise’s goal under a lot of constraints of planning condition. The method of traditional TFT-LCD material planning is sorting individual supply and demand first, and then locating them. The solution produced by this method is just feasible, not optimal. It can not satisfy enterprise’s demand goal at all to decrease productive cost. Due to, Genetic Algoritgm is needed a lot of computing time and cost to search solution. The computing time will increase exponentially by the size of problem, so we can not find the optimum quickly and efficiently. In this research, we will use constraint-based genetic algorithm that hybrid genetic algorithm and forward checking to apply for material planning problem in TFT-LCD industrial to maximize productivity.