Optimal distribution network configuration for the single-commodity, single-supplier, W-distribution point, N-retailer system
Abstract
This research examines single-commodity, single-supplier, N-retailer distribution systems that include any number of W possible non-stocking distribution points. The centrally-controlled system operates in a periodic-review base-stock replenishment environment required to attain a pre-specified customer service-level. Each of the N non-identical retailers supply normally distributed demand and is served directly from the supplier or through one of the W distribution points. Inventory costs are assessed on the safety stock and in-transit inventories. Transportation costs are assessed as a function of distance and total shipment weight. The per-unit transportation cost is a decreasing function of shipment weight. A constant per-unit cost is assessed for goods processed by any distribution point. Transit time, a surrogate for inventory leadtime, is a function of distance and also shipment weight. A primary objective of this research is to demonstrate the dependency between the inventory model decisions and the transportation and warehousing model decisions and the effect of this dependency on the total system expected cost of a distribution system. More specifically, it is desired to demonstrate that the optimization of a distribution system with respect to only the inventory costs without taking into account the effect of inventory routing decisions on the transportation and warehousing costs as well as the transit times (i.e., leadtimes) resulting from the routing decisions, will, in most cases, lead to a sub-optimal solution. Similarly, optimization with respect to only the transportation and warehousing costs while ignoring the impact of the transportation decisions on the inventory costs may also lead to a sub-optimal solution. The information gained from the examination of both the source and the magnitude of the dependency between the inventory, transportation, and warehousing considerations will be valuable in attaining another primary objective of this research which is the identification of the distribution system that minimizes the total expected cost per order cycle. In order to attain the second objective, a multi-phased heuristic is developed and tested. The results of the tests demonstrating the value of the heuristic and the insights provided by these tests are presented.
Degree
Ph.D.
Advisors
Sparrow, Purdue University.
Subject Area
Industrial engineering
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.