Large-scale, complex engineering systems, as for automotive manufacturing, often require significant capital investment and resources for systems configuration. Furthermore, these systems operate in environments that are constantly changing due to shifts in macroeconomic, market demand and regulations, which can significantly influence systems' performance. It is often very difficult or prohibitively expensive to change these engineering systems once they are in place. Thus, a critical question is how to design engineering systems so they can perform well under uncertainty. Conventional engineering practice often focuses on the expected value of future uncertainties, thus leaving the value of flexible designs unexplored. This research develops a new framework to design and plan large-scale and complex manufacturing systems for uncertainty. It couples a screening model to identify promising candidate solutions with an evaluation model to more extensively quantify the performance of identified solutions. The screening model adaptively explores a large decision space that is otherwise computationally intractable for conventional optimization approach. It integrates strategic and operational flexibility in a system to allow systematic consideration of multiple sources of flexibility with uncertainty. It provides a means to search the space for system's improvement by integrating the adaptive one-factor-at-a-time (OFAT) method with a Response Surface method and simulation-based linear optimization. The identified solution is then examined with Value at Risk and Gain chart and a statistics table.
展开▼
