| This dissertation focuses on the development of robust, context-independent representations for a wide class of product design problems. These representations permit the inclusion of both engineering and strategic considerations at the design stage, and, consequently, can be used not only as a part of a concurrent engineering effort, but also to evaluate the impact of strategic decisions. the AND/OR tree, we successively model increasingly complex interactions between design decisions. These models lead, in turn, to the development of efficient algorithms for solving a wide variety of multiobjective product design problems—we present applications of the modeling framework to the design of printed circuit board assemblies and microwave modules. By developing a framework for optimizing downstream metrics at the design stage, the dissertation moves beyond traditional concurrent engineering models, which are focused towards the evaluation of a given design with respect to the metrics and do not explicitly search the universe of design choices in order to determine optimal configurations. Furthermore, by incorporating supplier related issues at the design stage, it provides a first step towards an integrated product design-supply chain management system aimed at eliminating the design-strategic feasibility evaluation-redesign loop that is fast replacing the design-build-test-redesign loop of old. Next, we extend the models to incorporate demand considerations, and, in doing so, provide a starting point for the analysis of the interactions between manufacturing decisions and the marketing implications of those decisions. Finally, we examine the impact of supplier contracts on product design decisions, and thereby provide a first step towards the integration of the purchasing and the design functions within a manufacturing firm. The design of all of the above models and algorithms was driven by interactions with manufacturing and design engineers, developers, and managers at a Fortune 100 firm. Thus, the research, while motivated by the interesting and challenging underlying theoretical issues, was also focused on the development of models that could be easily implemented in a real-life design environment, and algorithms that could be feasibly embedded into existing decision support systems. We provide a brief description of an integrated system that was implemented in conjunction with the above mentioned firm. |