A period inventory routing problem and its implication in designing the distribution structure in a supply chain.
Item
-
Title
-
A period inventory routing problem and its implication in designing the distribution structure in a supply chain.
-
Identifier
-
AAI3127857
-
identifier
-
3127857
-
Creator
-
Chen, Hong Deng.
-
Contributor
-
Adviser: T. William Chien
-
Date
-
2004
-
Language
-
English
-
Publisher
-
City University of New York.
-
Subject
-
Business Administration, Management
-
Abstract
-
We have studied a period inventory routing problem (IRP) at the operational level where the supplier determines the replenishment policy for each retailer and the delivery routes per period, so as to minimize the long-term average cost on vehicle maintenance and dispatch, retailer inventory carrying, and vehicle routing. We have constructed three models. The first deterministic single-period capacitated vehicle routing model determines delivery routes and the fleet size. The second deterministic multi-period inventory routing model provides the fleet size, replenishment policy for each retailer and delivery routes per period. The third model extends the second one to include stochastic demand and stockout.;We develop three space-filling curve based heuristics to solve the above problems. The first algorithm develops a near optimal Traveling Salesman tour (TSP tour). The second algorithm partitions the TSP tour from the first algorithm into multiple vehicle routes according to retailers' demands and vehicle capacity. The third algorithm solves the multiple-period IRP and determines the fleet size, replenishment policy and delivery schedules. We then modify the third algorithm to handle stochastic demand and stockout.;We conduct experiments on randomly generated problem instances to demonstrate the effectiveness of our algorithms. We compare our solution with (1) a lower bound, (2) the solution of Clarke & Wright, (3) a performance estimator developed by Dannenbring (1977). Experimental results show that our algorithms are effective in providing near-optimal solutions to the studied problems.;To study the operational implication of our proposed solutions at the tactical level and to provide a systematic and quantitative approach to help decision makers gain insights on different aspects of their distribution system, we apply our algorithms to examine two strategies for designing the distribution structure in a supply chain: centralized system and decentralized system. Parametric experiments on product value, demand uncertainty, and service level indicate that high value product, high level of demand uncertainty, and high service level, generally favor the decentralized strategy, while the opposites prefer the centralized strategy. The experimental results also show that retailer density in a given area plays a significant role in determining the best supply chain strategy.
-
Type
-
dissertation
-
Source
-
PQT Legacy CUNY.xlsx
-
degree
-
Ph.D.
