Abstract—Manufacturing strategy comprises decision- making problems in terms of manufacturing practices to achieve manufacturing objectives through linkages of performance measurement. It is pervasively influential, long-term, and dynamic owing to the conformance with corporate strategy, business strategy, and marketing strategy. Therefore, there are many underlying sub-problems which can be expressed in a wide spectrum of forms. In this study, we will firstly define the spectrum of problem-structuring. After that, a decision analysis framework as a guide for modeling problems in different forms is introduced. Finally, a realistic problem is illustrated for discussions. I. INTRODUCTION HE manufacturing strategy is crucial for a manufacturing company to sustain competitive advantage. The meaning of manufacturing strategy is twofold. On one hand, manufacturing strategy must achieve corporate goals for sustainable competitive advantage. To link with corresponding actions, corporate goals are often decomposed into manufacturing priorities. By definition in the content model, manufacturing priorities are “a set of goals for manufacturing which is used to align the business strategy and market requirements with the manufacturing task” [1-3]. On the other hand, manufacturing strategy is responsible for making choices in terms of manufacturing functions. In particular, a pattern of decisions within certain categories affecting the ability of the manufacturing function should be made to meet the long-term objectives and the manufacturing task. The decision categories are usually divided into structural (related to long-term commitments and heavy investments) and infrastructural (related to support functions where changes may be incorporated in a shorter time perspective) ones, as proposed by Hayes and Wheelwright [4]. Additionally, manufacturing strategy is long-term as it takes an extended time horizon to plan, implement, and observe the outcomes. Finally, manufacturing strategy comprises a wide spectrum of activities that must conform to corporate strategy, business strategy, and marketing strategy. Commonly, a company seeks long-term profitability by maximizing revenue, the product of selling volume and unit price, and minimizing total cost. Indeed, the causal relationship between revenue and cost can be complicated. For example, a customer-driven company may try to minimize total cost so that it can offer wider range of prices Chen-Fu Chien is with the National Tsing Hua University, Hsinchu, Taiwan 300 ROC (corresponding author to provide phone: +886-3-574-2648; e-mail: cfchien@mx.nthu.edu.tw/cfchiena@tsmc.com). Jei-Zheng Wu is with the National Tsing Hua University, Hsinchu, Taiwan 300 ROC (e-mail: d907804@oz.nthu.edu.tw). for customers. Consequently, the company may lower prices to gain more selling volume. Thereby, profit could be increased with proper price setting. However, as previously discussed, multiple factors such as quality and delivery may also influence price other than cost. Competitors in the same market may thus find their niches [5]. In a complex version of causality between revenue and cost, manufacturing objectives is the immediate source of revenue and manufacturing practices is the source of cost. Moreover investments in expensive equipment are not objectives but means to satisfy demand. However, this means must be able to link with the manufacturing objectives immediately or indirectly so that effort will not merely produce cost. Although non-financial indexes has been emphasized for measuring organization performances [6], little study has been done in structuring manufacturing strategy with causal relationships. Hence, this study aims to introduce a problem-structuring technique to structure the causal relationships in manufacturing strategy. II. THE CONCEPTUAL PROBLEM We firstly shed light on the extent of problems being discussed. Without loss of generality, a decision-making process comprises a real system, a conceptual problem, and execution as shown in Fig. 1. The paradigm of model development process is originally proposed by Sargent [7] for simulation model development which can be extended for the development of general decision-making models. Since a simulation model requires unambiguous definitions of objectives and relationships among entities, the paradigm should be extended so that more decision-making considerations can be included. In the paradigm, a real problem is an idea, situation, policy, phenomena, or a sub-problem to be modeled under a manufacturing strategy. For example, a company may encounter an outsourcing problem for choosing a set of qualified vendors and allocating orders among them; or a facility may wonder how to conduct preventive maintenance schedules over a number of machines; or a top manager would seek how expensive equipment be measured to improve productivity and to identify root causes of failures. On the other hand, the conceptual problem, corresponding to a conceptual model, is a mimic of the real problem in terms of mathematical, logical, verbal, or graphical representation. Many conventional problems have been discussed from multiple disciplines in academia, e.g., vendor selection problems [8-10], diagnosis and fault detection problems [11-12], site-imbalance and multi-chamber problems [13-14], Structuring Manufacturing Strategy Chien-Fu Chien, Member, IEEE, and Jei-Zheng Wu T Proceedings of the 3rd Annual IEEE Conference on Automation Science and Engineering Scottsdale, AZ, USA, Sept 22-25, 2007 SuRP-B03.3 1-4244-1154-8/07/$25.00 ©2007 IEEE. 265