9622647 Ehrhardt This research deals with finished goods management models for just-in-time (JIT) production. Several different formulations of the dynamic decision making environment will be considered, eventually emphasizing the infinite-horizon version of each model. In each case the general objectives will be essentially the same: (a) understand how optimal policies and costs depend on order size and schedule stability (using theoretical analysis of dynamic programming models), (b) develop effective heuristic policies that have modest requirements for demand information and computational effort (using theoretical and computational studies of dynamic programming models) and (c) investigate system behavior when demand parameters are estimated by periodically revised statistical estimates (using discrete-event digital simulation models). This research should result in mathematical models sufficiently developed so that implementation tactics can be developed for facilitating the management of the interface between a JIT manufacturing process and unpredictable external demand for finished goods. Implementable computer software, ready for field testing, should be in place.
