Economic models for determining excess inventory levels for the case of non-constant demand
manufacturing philosophy as practiced by many Japanese firms. Many advantages accrue to the manufacturing process when JIT is implemented. Less clear,...
manufacturing philosophy as practiced by many Japanese firms. Many advantages accrue to the manufacturing process when JIT is implemented. Less clear, however, is the impact on the firm’s logistics system. The objective of this research is to identify the impact of a JIT strategy on a firm’s logistics system, and to identify operating conditions where JIT implementation does not adversely affect the logistics system. Annual logistics cost of a JIT strategy is compared to that of a theoretical optimal strategy and consolidation of inbound shipments is considered as a means of achieving transportation economies of scale in a JIT strategy. Fifteen parameters are used to identify inventory and transportation conditions of manufacturers and vendors. Using simulated data representing alternative business scenarios, paired t-tests are used to identify differences in annual costs between the strategies. Once the annual logistics cost impact of a JIT strategy is established, correlations are used to identify which cost components, transportation, inventory carrying, ordering, or expected stockout, are related to the differences between the annual logistics costs of the strategies. Those cost components that are correlated with the difference in costs between the strategies are then further partitioned to determine the cause of the cost differences. Next, correlations are used to identify the relationships between the fifteen inventory and transportation characteristics and the cost differences between the strategies. This will provide information for logistics decision-makers considering the potential impact of a JIT strategy. Finally, regression is used to develop equations that predict the effects of the fifteen inventory and transportation factors on the annual logistics cost differences between the strategies. (Order Number DA8913449, October 1989.)
Economic Models for Determining Excess Inventory Levels for the Case of Non-Constant Demand RICHARD ALAN TOELLE, PH.D. THE UNIVERSITY OF OKLAHOMA, 1988, 264 PP. MAJOR PROFESSOR:RICHARD J. TERSINE
This dissertation develops and compares several approaches to determining optimal inventory liquidation policies when future demand is not expected to be constant. The role of economic modeling in the detection, disposition, and control of excess inventories is also discussed. Models of the excess inventory liquidation decision are developed for use under a variety of circumstances. Both present-value adjusted and unadjusted models are developed. The analytical models view time as a continuous variable and lead to closed-form optimality conditions. A dynamic programming algorithm and two heuristic approaches are devised for use in an MRP environment. A model is developed which views future demand as a random variable. The time of demand for the marginal unit retained is shown to be an order statistic. It is also shown that the optimum policy can be determined using a deterministic method. However, a measure of the degree of risk associated with the optimum policy is provided by the random demand model. Finally, the performance of three generic approaches to liquidation decision modeling is compared using simulation. The performance of the dynamic programming algorithm is shown to be superior when there is no forecast error. The three methods perform equally well when there is significant forecast error. (Order Number DA8913787, October 1989.)