隨選目標之智慧型排程與產品管理系統之研發

A System Development of Intelligent Scheduling under Selected Objective and Products Management

本研究針對零件代工製造商開發一套產品與排程管理系統，除了基本的進貨、銷貨及存貨管理外，更加入了智慧型排程系統，可根據不同的訂單狀況選擇不同的績效指標。研發過程中，本研究採取階段式比較方法尋找最佳排程結果。首先，我們沿用一般零件代工製造商常用的先到先服務法，再依一般的生產排程方法，即最短作業時間、最早交期時間和關鍵比值法，進行第一階段之排程作業，若其最佳排程結果仍有延遲現象，則系統自動進入第二階段，亦即採用本研究所開發之改良型基因演算法進行排程結果的改善與優化。本系統可自動產生最佳的排程結果，且能有效改善原先排程結果之績效。本研究將智慧型排程方法崁入管理資訊系統中，經測試結果發現，確實能提升管理效能與排程績效。

This study examined the OEM manufacturer of parts to develop aproducts management and production scheduling system. Besides purchase,sales and inventory management, we proposed an intelligentscheduling system that search for an optimal scheduling result depend ondifferent performance indicators. The searching steps are as follows. Atfirst, the first-in/first-out method was taken, and then shortest processingtime, earliest due date and critical ratio methods were followed. If theresults were not good enough, then a revised genetic algorithm method wasapplied to improve the scheduling results. In general, the proposedsystem can improve the scheduling performance more efficiently than thetraditional scheduling approach.