標題: 多工段基因導向派工法則之研究Genetic-Based Dispatching Algorithms for A Multi-Stage Textile Manufacturing System 作者: 陳瑩芝Y-C Edna Chen巫木誠許錫美Muh-Cherng WuHsi-Mei Hsu工業工程與管理學系 關鍵字: 格子布紡織業;派工演算法;多工段生產系統;成組交貨;基因演算法;yarn-dyed textile industry;dispatching algorithms;multi-stage manufacturing system;group delivery;genetic algorithms 公開日期: 1999 摘要: 因流行時尚的多變，格子布紡織業面臨瞬息萬變的需求趨勢，紡織業者需對客戶提供較短交期的服務以提升其競爭力。格子布紡織業為純客製化的生產系統，其生產過程主要分為以下五個工段: 染前整經→染色→染後整經→漿紗→織布為一多工段生產系統。因染色與織布工段是整個生產過程中的瓶頸，這兩個工段的派工方法對訂單是否能達交有極大的影響。 本研究以總訂單延遲天數最小為目標，以基因演算法為基礎，提出一套求解多工段的派工演算法。本文所提之多工段的派工演算法包含三個模組，分別為後推模組、前推模組與調整模組。重覆執行三個模組直到產生一組滿意的解才停止。 本文以一實際紡織廠為例說明多工段派工演算法，該演算法可快速的求出近似最佳解的可行解。Due to dynamic change of clothes fashion, short cycle time service has been a very important criterion in the yarn-dyed textile industry for getting customer orders. Yarn-dyed textile is a special kind of textile and its production system in general is a make-to-order (MTO) system. Of the five manufacturing stages in the production of yarn-dyed textile, dyeing and weaving processes capture most parts of the total cycle time. Dispatching algorithms in these two processes, dyeing and weaving, are therefore very much important in the reduction of total cycle time. This research proposes a series of dispatching algorithms for sloving dispatching problem of a multi-stage manufacturing system with the objective of minimizing total tardiness of orders. These dispatching algorithms are both genetic-based, and are mutually linked through uses of a backward inference mechanism, a forward inference mechanism and an adaptive inference module. In the backward inference module, given customer due dates of orders, we use the weaving dispatching algorithm for determining the targeted due date of the dyeing process by assuming the capacity of dyeing is infinit. Then, the targeted due date for dyeing is used in the dyeing dispatching algiorithm for determining the "best-possible" releasing schedule of the dyeing process, which may not be feasible in achieving the targeted due date of dyeing due to infinit capacity assumption. The forward inference mechanism is then applied; that is, the "best-possible" releasing schedule of the dyeing process is used as input to generate the releasing schedule of the weaving process. From the adaptive inference module, a new dispatching sequence for the weaving process is subsequently generated by considering two constraints: due date of orders and releasing schedules of orders. The backward and forward inference mechanism are alternatively applied until a satisfactory solution is achieved. The proposed approach has been compared with the scheduling method used in a real factory, and has shown better results. 1.1研究背景與動機…………………………………………………………………1 1.2格子布紡織生產流程簡介………………………………………………………2 1.3研究目的與範圍…………………………………………………………………7 1.4研究限制…………………………………………………..……………………7 1.5論文架構…………………………………………………..……………………7 第二章 文獻回顧…………………………………………………………………..9 2.1多工段生產系統………………………………………………………….……9 2.2紡織業中的排程派工應用….…………………….……………………….…9 2.3 Sequence Dependent.….….……………..……..……………………….13 2.4基因演算法..….……………………..………………………….…………14 第三章 問題分析與派工法則之發展…………………………….………………17 3.1問題定義………..…………………………………………………………….17 3.2解題構想……..……………………………………………………………….18 3.3格子布產業派工演算法之發展……………………………………………….22 第四章 構建基因為底的派工法則演算法……………………………………….28 4.1以基因為底的織布工段後推模組.……………………………………………28 4.1.1線性規劃模式…..………..…………………………………………28 4.1.2以基因為底的織布工段後推模組…..………………………………31 4.2以基因為底的染色工段後推模組…………………………….………………44 4.3以基因為底的織布工段前推模組………………………………………….…50 4.4以基因為底的調整模組………..……………..………………………….…52 4.4.1調整模組……………..………………………………………………52 第五章 實例驗証……..………………………………………………………….55 5.1織布工段後推模組(一)..……….….…………………………………….55 5.2染色工段後推模組(一)………………………………………………………56 5.3織布工段前推模組(一)………………………………………………………57 5.4調整模組(一)………..…………..…………………………………………58 5.5染色工段後推模組(二)………………………………………………………58 5.6織布工段前推模組(二)………………………………………………………58 5.7實例驗証求解流程的總結….…………..…….……………………………59 第六章 結論與未來研究方向……………………………………………………62 6.1結論………………………..…………………………………………………62 6.2未來研究方向……………..…………………………………………………63 參考文獻.……………………………………………………………………………64 URI: http://140.113.39.130/cdrfb3/record/nctu/#NT880031037http://hdl.handle.net/11536/65195 Appears in Collections: Thesis