CONTRIBUTIONS TO DISTRIBUTION SYSTEMS THEORY IN MULTIPLE OWNER, MULTIPLE PRODUCT ENVIRONMENTS

This thesis analyzes two important aspects of the distribution logistics system: (1) Different locations may have different owners (e.g., dealers and suppliers) having different priorities with respect to cost and service; (2) The multiple product nature of these systems suggests that strategic management of the distribution function is more effectively achieved by dealing with categories of parts rather than dealing with single parts. We first formulate a comprehensive model for analyzing the behavior of dealers, in particular, how they prioritize customers and how they select sources for replenishment. After considering the associated Markov process model, an approximate, renewal based model is derived. This approximation is used to develop a heuristic which minimizes expected cost subject to fill rate service level constraint. Extensive numerical testing of both the accuracy of the approximation and the performance of the heuristic with respect to the true optimal solution is included. A complete sensitivity analysis is carried out to understand the trade-offs and available alternatives faced by the dealers when the operational parameters are changed. We then introduce a clustering methodology for assigning stock keeping units (SKU) into groups for planning and control of production/distribution systems. The methodology was developed to support modelling and analysis of production/distribution strategies. The main departure of this analysis from earlier work is: (1) the approach can handle any combination of item attribute information that is important for managerial purposes; (2) management's preferences for groupings based on operational performance can be accommodated; (3) statistical discrimination criteria are considered, and (4) group definition reflects the application context, e.g. control policies based on group membership. The approach generates groups which we refer to as Operations Related Groups (ORGs). An application based on a study of inventory policies for an automobile spare parts distribution is also presented. The performance relative to the commonly used ABC method is especially significant.