在緊急訂單與延遲成本考量下零工型排程問題派工法之硏究

中文摘要

緊急訂單的處理爲零工型製造環境中常見且相當棘手的問題。由於其交期緊迫度與未滿足交期所衍生之延遲成本均較一般訂單爲高，因此於現場加工時需給予其適當之優先權。然而若未能適當處理，除了可能造成緊急訂單的延遲外，另一方面亦可能影響產能的分配與一般訂單之加工，進而造成更大之延遲成本。因此，本研究針對零工型排程問題所需面對之緊急訂單處理方式，考慮訂單等級與交期以及產出平衡之槪念，提出一整體考量之派工法，藉由模擬與基因演算法之運作，配合交期緊迫度、單位時間延遲成本、各製程完工量、機台負荷等資訊之收集與應用進行派工決策，以最小化整體之總延遲成本。經由模擬實驗與統計檢定證實本研究所提出之派工方式不論在總延遲成本或交期達成率上，均較SPT、SRPT、FIFO、EDD、CR、SLACK此 六種常見之派工法則有更好的表現。

Abstract

Rush order management is a common but troublesome problem in the job shop manufacturing environment. Due to its tight due-date and high delay cost, it is a common practice to give a rush order a higher priority over normal orders in the shop-floor dispatching decision. To do so, however, it may cause more late normal orders and incur heavier delay costs. This study proposes a dispatching procedure incorporating simulation and genetic algorithm for the job shop environment with order rating and due-date considerations. The concept of output smoothing is introduced. The objective function is to minimize the total delay costs. Through simulation experiments and statistical tests, the proposed dispatching procedure gives the best due-date performance and the minimal total delay costs compared with those of SPT, SRPT, FIFO, EDD, CR and SLACK.

關鍵字:

• 緊急訂單
• 延遲成本
• 派工
• 基因演算法
• 模擬

Key Words:

• rush order
• delay cost
• dispatching
• genetic algorithm
• simulation