SINGLE LEVEL AND MULTI-LEVEL, LOT SIZING STRATEGIES IN MATERIAL REQUIREMENTS PLANNING SYSTEMS (OPERATIONS MANAGEMENT)

The area of Material Requirements Planning (MRP) is receiving major emphasis in the management of operations due to the financial benefits of efficient deployment of inventory investment. The issue of lot sizing rule has been one of the most heavily addressed topics in the MRP area. Prior research in lot sizing has dealt with single level or multi-level problems with a single parent product structure. However, real world problems frequently involve both multiple levels and multiple parents.;We formulate a mixed integer programming model for the multi-level lot sizing problem to facilitate the development of a multi-level lot sizing algorithm. An examination of sample solutions obtained by mixed integer programming with the ones obtained by a sequential application of any single level lot sizing rules can provide a pattern which would be useful in developing a recursion algorithm. Exploiting the characteristic differences among these solutions, a recursion algorithm is developed.;Three popular lot sizing rules are modified for the single level, and multi-level problem as well. The three modified lot sizing rules operate on the part period accumulation principle besides their original mechanism.;To evaluate these proposed single level and multi-level lot sizing heuristics we develop a computer simulation model of an MRP system and design experiments designed to consider different demand patterns, different degrees of commonality and different numbers of levels in product structures. Evaluation criteria include inventory costs and computing time requirements.;To develop a multi-level lot sizing rule, the dependent demand relationship between parents and components and common usage of certain items in producing different finished products should be considered in some fashion. The objective of this research is to develop multi-level lot sizing rules that are simple to understand and can be implemented with ease on an MRP system. This dissertation proposes a new approach to multi-level lot sizing for MRP systems.