The inventory system considered in this work is single echelon, N-location and continuous review in which complete pooling of stock is permitted among the locations. A model is developed for slow moving, expensive items that are common to two or more locations. The model developed focuses on parts that have a lumpy demand pattern; that is, the number of units demanded at each demand occasion is a random variable. The motivating case for this research has been a real-life situation presented by a large utility company having 29 power generating plants in five southeastern states. Data analysis from the company reveals that there are a substantial number of parts that have low usage, are expensive, and have a lumpy demand pattern. At present, each location operates independently and maintains enough stock to meet its own requirements. Transshipments take place between locations whenever there is an emergency requirement for a part, but no explicit consideration is given to this effect while deciding on the inventory control policies at different locations. A model, using analytical and simulation techniques, is developed for this situation. Limited experimentation, performed on a set of problems, shows an average savings of about 31% as a result of pooling.
