Extensions To a Model For Tactical Planning In a Job Shop Environment

We develop three extensions to a previously published model which assists managerial planning in a multi-sector job shop [Graves, 1986]. This model provides insight into the trade-off between two aspects of system behavior, production smoothing and size of work-in-process inventories, which will be affected by management's choice of production lead times. For all three extensions we provide numerical examples of application. First, we develop a model of a production release rule which coordinates demand forecasts over a planning horizon with knowlege of the projected completion times of items currently in work-in-process inventory in order to determine the number of items to start into production. This release rule model is necessary in order to apply Graves' model to a production line which produces in response to demand forecasts rather than in response to order arrivals, as is typical of a job shop. Second, we develop two measures of the service level of a shop and show how these are affected by the choice of production lead times. The service measures considered are the probability that demand during a random period is met and average length of failure runs (by a failure run we mean several periods in succession during which the facility fails to meet demand). We discuss how management can evaluate the impact of lead time choice on the expanded set of trade-offs among service measures, production smoothing, and size of inventories. Third, we model the dynamic behavior of a shop in response to various types of change in demand. The changes considered are a one-period increase, permanent increase, linear growth, and cyclical demand. We show how to assess the time the system will take to achieve its new equilibrium and the nature of the path followed during the transition and how these responses depend upon lead time choices. These three extensions should enhance the usefulness of Graves' model by widening the scope of facilities to which it may be applied and by extending knowledge about the impact of lead time choices on additional important aspects of system behavior. Thesis Supervisor: Dr. Stephen C. Graves Title: Associate Professor of Management Science