A Bayesian Decision Model for Process Development Based on Cost and Quality Outcomes Abstract Automobile manufacturers currently use two approaches to determine when to accept a production tool: Conformance to Specifications (the tool produces products within design specifications) and Functional Build (the tool produces products that function as the customer expects). This project will develop an analytic framework for deciding which of these is appropriate, and the associated policies for optimal tool testing before using the tool in a production setting. The objective of this research is to provide manufacturers with a decision tool, based on available cost of quality data from existing accounting systems, that will reduce total manufacturing costs. This is in contrast to current accounting practices, which drive separate cost centers within a large organization to suboptimize performance relative to their own budget. Thus, the decision models developed by this research will cut across organizational boundaries to provide manufacturing firms with optimal financial and quality results. The first phase of the research will benchmark die buyoff. This will allow identification of the metrics (financial and quality) currently used to track process development. The output of this initial research will be a set of die construction and product quality performance measures that will form the framework of the decision model. The second phase of the research will focus on the development of methods for obtaining and tracking the performance measures previously identified. The third phase will represent the construction of models, combining methods from decision analysis, Bayesian statistics, and accounting systems, to focus on developing the methodology that optimizes the decision process for the entire organization.
