格子布染色工段派工方法之探討

Abstract

本研究對格子布紡織廠染色工段的派工提出三個啟發式演算法，並以某一實際格子布紡織廠的資料，驗證這些演算法的績效。格子布的製造過程複雜，一個訂單包含許多不同樣式的格子布，每個樣式通常由好幾種顏色的紗織成。這些染完顏色的紗是整經後的原紗因型態或重量不同，經由生產規劃會被指派到不同的缸型進行染色作業的產出。染完後如果該色紗是屬經紗則送到漿紗工段，漿完後送到織布工段；如果是屬緯紗，則送到織布工段。屬同一樣式的所有顏色的紗全部到達後，才可進行該樣式格子布的織布作業。屬同一案的布料需全部織完後，才可算完工。這種成組齊一完工的特性，造成染工段在各缸型決定待染工單的派工優序時，需考慮染色工單間的關係。因此本研究考慮染缸作業『順序相依的設置時間』與『成組齊一產出』的特性，以最小化所有案總延遲時間目標，發展三套適合紡織業中解決染色工段派工的啟發式演算法（分別為模擬退火法，基因演算法以及禁忌搜尋法），供管理者作參考以及應用，研究結果顯示這三種啟發式演算法的績效皆比傳統的派工方法優異。其中模擬退火法及禁忌搜尋法在所有案總延遲時間及搜尋時間上，有較佳績效。

This research proposes three heuristic dispatching algorithms（Simulated annealing ,Genetic algorithms ,Tabu search）for the dyeing process in a yarn-dyed textile factory. The performances of these algorithms are verified by a real case. Special characteristics of the yarn-dyed textile industry, that have been taken into account in the development of the dispatching algorithms for dyeing process, are the following: (1) the set-up times of the dyeing machine are dependent on the sequence of jobs on the machine; (2) the tardiness of a job is dependent on the other jobs which are included in a particular pattern of customer order. The second characteristic is called “grouping delivery” The performances of three proposed algorithms have been compared with those of the traditional dispatching methods (EDD, SPT, CR), and have shown better results. Among the three proposed heuristic dispatching algorithms, Simulated annealing and Tabu search have better performances and have less computation time than Genetic Algorithms.