多組件產品供應商的選擇模式

A Supplier Selection Model for A Product Containing Multiple Components in Series Reliability Structure

吳佳蓉(Chia-Jung Wu)

指導教授 ： 徐旭昇

元智大學/工程學院/工業工程與管理學系/碩士(2000年)

摘要

本文發展一貝氏向後選取的模式,針對兩組件可靠度串聯結構產品做研究,以最小期望總成本為目標,假設組件檢驗存在誤差但產品的檢驗則無誤、最終的不良產品是裝配好的,且無法再重新拆解組裝。考慮的成本參數包括組件購買單價、檢驗成本、以及產品失敗成本。當一組件的檢驗結果為不良品時,將展開遞補行動;從備批內連續抽驗組件,直到檢驗結果為良品的組件出現替換為止;備批內組件的品質與購批內的一致。此一額外產生的遞補成本,可在契約內規定由廠商自付。在期望成本值的計算方面,則使用到最佳樣本數與模式參數的相關性,以及品質的事前分配。其中模式參數考慮到組件品質、檢驗水準以及上述的成本參數。而品質則服從Beta分配。

關鍵字

抽樣計劃 ； 貝氏決策分析 ； 決策樹

並列摘要

In this paper we develop a Bayesian Vendor Selection model for a product containing two components in reliability series structure. Our model assumes that there can be inspection errors while inspecting components but not at final product stage. The model takes into accounts component quality, the purchase cost, inspection cost, product failure cost with minimizing the total cost as the model objective. A failed product can not be disassembled and reassembled. When a component is reported “bad” during inspection, a series of inspections will be done on the spare lot containing the components of the same quality until a reported “good” component is found. Such extra cost can be charged to the producer, depending on the contract. Computational results using the Beta priors for component quality are presented and the relationships between the optimum sample size and the model parameters such as component quality, inspection quality, and the above costs are analyzed.

並列關鍵字

Inspection sampling plans ； Bayesian decision analysis ； Decision trees

參考文獻

[2] 陳金平 「多組件產品非同步抽樣貝氏檢驗計劃之設計」

[1] 林鴻欽 「貝氏檢驗計劃的設計」

[3] 蘇小惠 「貝式方法解析考量檢驗誤差之戴明成本模式」

Position.〞 Massachusetts Institute of Technology, Center

Discussion of Acceptance Sampling Procedures.〞In:A.P.