可拆解再裝配產品之貝氏檢驗模式

摘要

本論文旨在提供製造商兩計量模式適用於其產品屬於可拆解再裝配狀況時，該如何進行進料抽樣檢驗及如何選擇組件供應商。模式一適用於產品屬於多組件串聯結構，模式二則加入兩種型態檢驗誤差考量但限制適用於單一組件結構。這兩種檢驗模式皆以貝氏決策理論爲架構基礎之修正檢驗與屬性抽樣。模式成本考慮的範圍涵蓋組件之檢驗成本、不良品所發生之額外檢驗成本及其所導致產品失效成本，而以最小期望成本爲目標。我們首先對組件品質不設定分配之下，對檢驗模式進行理論分析，獲致一些結果。然後在組件品質假設爲Beta分配之下做實驗分析，綜合理論與實驗析結果撰寫上述兩模式之計量快捷程式，以符合實際運作時的方便性與時效性。

Abstract

The purpose of this paper is to present two inspection sampling models for products that can be disassemble and reassemble when they test to be failures. The first model can be applied to products having a series structure of several identical components. The second model studies the case of single component products in consideration of possible inspection errors. Both models can also be used to select the best supplier when the total cost of the product is the main concern of the producer. Many studies have been done in regard to products having single component in series structure, but very few in regard to products having multiple components in series structure. Some theoretical results in the model for general prior distributions are derived. In addition, computational results with Beta distributions are presented so as to help readers understand the relationships among the inspection costs, the optimum sample sizes, and the total cost of the products. The results also provide a producer with a guidance for taking actions on an incoming lot and a good reference for selecting suppliers from an economical point of view.

關鍵詞:

• 檢驗計劃
• 貝氏決策理論

Key words:

• Inspection Sampling
• Bayesian Decision Analysis
• Decision Trees