OPTIMAL LOT-SIZING PROCEDURES FOR MULTILEVEL STRUCTURES

Since the introduction of Material Requirements Planning (MRP), several traditional approaches become inadequate for the new assumptions and relations involved in these systems. This research has concentrated on one specific area that requires new approaches and that provides opportunity for quantitative methodology applications. The classical single-level optimal lot sizing procedure, i.e., the Wagner-Whitin algorithm, does not provide optimal solutions for the multi-level problem. One of the main purposes of this research was to understand the behavior of the optimal lot sizes in multi-level product structures, especially in parent-component relations. One approach has been developed to solve the multi-level optimal lot sizing problem. It is a new implicity enumeration procedure which makes use of a method to obtain K best order policies for single items formerly developed in this study. The huge computer time required by this procedure, encouraged the formulation of an approximation method efficient enough to be used in any practical situation. Special cases were studied and the corresponding characteristics used to develop the method. In an experimental with 5 items product structures and 12 period time horizons, this procedure produced an average total inventory cost (setup plus carrying costs) only 0.09 % higher than the optimal solution. The computer times were centesimals of seconds compared with hundreds of seconds needed to determine the corresponding optimals.