| Manufacturing capacity planning is a complex problem for many reasons, including constrained budgets, high levels of capital requirements to acquire expensive equipment, the need to produce sophisticated and customized products in a single manufacturing site, and uncertain demand streams.; This research develops methodologies for equipment selection for agile manufacturing in supply chains, and steady state capacity allocation to each operation. The study includes uncertain demand realizations, manufacturing multiple products in a single site, using multi-stage processes, and equipment with multi-process capability. Machine rates can be dissimilar for each operation in each equipment type. The modeling is done at the shop floor level.; Capacity planning under these conditions is a combinatorial problem. This research develops three models which are shown to work for real life sized systems: First, a Monte Carlo optimization model which characterizes the required tools based on the demand properties and the relations among technical variables. Second, a Set Covering model for minimizing investment requirements while generating near optimal configurations. Third, a Maximum Entropy model suited for the highest level of uncertain demand and complex configurations, which maximizes the number of efficient alternative routes for the product to advance through the manufacturing system. All three methods make use of a transformation of product-based queueing networks into operation-based queueing networks.; Complete numerical examples are developed for a clear understanding of the methods. The three models are validated in the diffusion area of a semiconductor manufacturing company. Finally, conclusions and future research are presented. |
