A distribution network design model based on network splitting
Metadata[+] Show full item record
[ACCESS RESTRICTED TO THE UNIVERSITY OF MISSOURI AT AUTHOR'S REQUEST.] In today's market place, logistics have become more important than ever. How to devise an efficient delivery network with low cost has attracted a significant amount of research over the years. The problem is called the Vehicle Routing Problem in literature, which is an extension of the Traveling Salesman Problem. Recently, a new branch of the classical Vehicle Routing Problem was proposed called the Split Demand Vehicle Routing Problem. It introduces improvements to the Vehicle Routing Problem. In this paper, a new approach to the Vehicle Routing Problem is developed in terms of network splitting. The network is split into several layers, and then solutions to the Vehicle Routing Problem at different layers are used to constitute the final solution. The routing problem at each layer is modeled as a linear programming model and solved by Gurobi 4.0.1. Then in order to find a good way to split the network, a Genetic Algorithm is implemented to determine how to assign the demand of sites at different levels of the network. A real case study was used to test the model. Based on the case study the advantages of this model are discussed. Next a larger case was examined in order to show the advantage of splitting a network compared with a model without it, plus ways to split the network are explored. Finally, future research directions are summarized.
Access to files is limited to the University of Missouri--Columbia.