- 學位論文
基因演算法於多目標全球成衣報價之應用
Application of Multi-Objective Genetic Algorithm to Quote in Global Garment Industry
摘要
在全球化的競爭市場中，企業為了要維持競爭優勢，必須要能快速地回應顧客的報價需求。由於傳統的人工報價過程非常地繁瑣，往往涉及公司多部門、跨區域的整合，因而需耗費多時才能完成報價。而且報價單的內容還必須與客戶不斷地溝通與修改，若僅透過傳統人工的方式處理，容易受到決策時間太短或決策者經驗不足等因素所影響，而產生錯誤的報價或是喪失訂單，造成公司的重大損失。 業界對於產品報價考量的目標很多，且目標間彼此往往互相衝突，為達到最佳化評估的目的，過去的研究常假設為單一目標或是給予各個目標不同的權重值，再以單目標的方式求解最佳值。這樣的方法所產生的結果常與最終目標有所偏差，其求解的過程亦相當繁雜，決策者往往需具備一定程度的專業能力與數學技巧，對於實際業者而言，並不適用。 為輔助業者解決報價的問題，並處理傳統最佳化方法之限制與瓶頸，本研究首先藉由與業者實際訪談，整理並規劃現今業者之報價流程，進而提出一報價機制，其包含了人工處理及電腦計算的部份，在業務人員接到訂單需求後，營運中心可先透過人力的方式，根據營運目標，並依照需求剔除不適合的生產廠區。其次，透過多目標基因演算法，找出柏拉圖解，將工單分配至合適的廠區，並計算出其成本及交期，進而制定價格。 研究結果顯示，在極短的時間之內，運用多目標基因演算法可求得多組不同的柏拉圖解，以提供業者在制定報價決策之參考依據。且其運算結果亦非常接近暴力搜尋法找到的結果，即代表本研究所產生的結果具備一定程度的準確率。
並列摘要
In global markets, it is very crucial for companies to enhance competitive advantages to achieve quick response quote demand from customers. The quoting process is difficult and complex. The traditional methods in solving this problem was that decisions were always made by senior managers’ own experiences in garment industry. Some mistakes may be made while the decisions are made in little time. Thus, some scholars suggested Genetic Algorithm (GA), which was adopted to analyze orders and give suitable results, to assist companies in decision making. However, this method may not be suitable for solving real situations since previous research employed single-objective planning problem to present real situations, whereas many of real situations are multi-objective planning problems. Hence, the aim of this study is to propose a quote mechanism and find the suitable analytic tool for multi-objective planning problems. The quote mechanism proposed by this study can be separated into two parts. First, when receiving order demands and sending demands to the operation office through Internet. In terms of own experiences and limitation of orders, unsuitable factories can be eliminated quickly for some important orders. Those orders are then allocated into suitable factories by computers. After interviewing with senior managers, there are two main operation objectives including “Minimum Cost” and “Minimum Makespan”. To analyze multi-objective planning problems, Multi-Objective Genetic Algorithm (MOGA) was adopted to achieve the main purposes of this study. The results indicate that the mechanism can assist users to find some non-inferior solutions in few seconds. In addition, the results are quite comparable to those by Brute-Force Search. Furthermore, this also explains that the results yielding from this study possess good predictive ability.