- 學位論文
Developing a decision support system to solve the problem of using surplus materials based on genetic algorithms
Developing a decision support system to solve the problem of using surplus materials based on genetic algorithms
摘要
製造業需要裁剪原料來生產成品時會有餘料或廢料產生，例如金屬類(模具、鋼條與鋁條等)。廢料是指該物料無法於製程中進行加工或做為原料使用，一般均作為下腳料賣給回收業者，餘料則是可以再利用生產。然而，生產不同的產品所產生的餘料格式和類別都不盡相同。因此廠商應先有效率地消耗餘料庫存，將生產過程中所產生的廢料降到最少，以節省原料成本，此亦即本研究所探討的核心問題。 本研究以啟發式方法來規劃餘料消耗的問題，並透過基因演算法求取近似最佳解。除此之外，本研究依據此方法，設計了最小配對法(minimum matching slack)與非配對交配方法(non-match crossover)。為了驗證這些方法的可行性與及優劣，將利用電腦程式語言C++開發一個決策支援系統，來有效協助基因演算法運作的過程；並透過系統實驗不同的鋁門窗生產訂單環境，來驗證所提方法的可行性，最後提出建議作為製造業者解決餘料消耗問題之參考。
並列摘要
The manufactured products will be surplus materials left. There are usually some surplus materials from any manufacturing process, and when the items being produced come in many types and styles, this makes the use of such materials more difficult. This presents a significant management problem for manufacturing firms, and thus the current study examines the issue of consuming surplus materials, the so-called macos-match problem, in the most efficient manner. This study proposes a heuristics algorithm that is combined with genetic algorithm to solve the problem of matching surplus materials to orders. In order to verify the feasibility of this method, a decision support system is developed using C++ language. This system is then applied in different experimental environments in order to compare three methods - non-match crossover, minimum matching slack and genetic algorithms. The method presented in this work is thus of considerable value to managers of manufacturing firms.