紡織廠染色工段產能分配模式之構建

Abstract

染色工段的排程規劃對訂單的達交率有很大的影響。目前業界接到訂單後，即依設計部門的判斷給定各訂單在染工段之製造工單的缸型，因缺乏整體規劃，造成染缸缸型的產能分配不均，結果使得各工單在染工段的生產週期時間過長，進而影響訂單的達交。本研究考慮染工段各製造工單可使用的缸型集合，最小化染完規劃週期內所有製造工單的時間為目標，構建缸型分配的整數線性規劃模式，以決定各製造工單的缸型。因該缸型分配模式的解空間過大，本研究利用基因演算法及CPLEX軟體，提出一套求解方法。 經由某一實際工廠的資料模擬驗證，在染工段工單之生產週期時間的平均值與標準差之生產績效指標上，利用本研究所提之缸型分配模式，指定各製造工單的缸型，顯著改善業界由設計單位指派製造工單缸型的績效。

This research presents an integer programming model to determine the type of dye machine for manufacturing orders on a weekly basis. For solving the large integer programming model, we suggest a heuristic method by using genetic algorithm(GA). Each solution is encoded as a string. Based on the solution represented as a string, we can find the numbers of manufacturing orders to assign to each type of dye machines. Then the numbers of manufacturing orders to assign to each type of dye machines are set as constraints added to the proposed integer programming for reducing the solution space. Software CPLEX is applied to solving the reduced integer programming model. Finally, a real case is used to illustrate the proposed approach. And a simulation model for this real case is applied to define the performances of cycle time in dyeing process for using the approach proposed in this research. The results show that the mean cycle time and the standard deviation of cycle time are significantly improved.