Electronic Theses and Dissertations

Identifier

6294

Author

Yuan Zhou

Date

2018

Date of Award

1-1-2018

Document Type

Dissertation

Degree Name

Doctor of Philosophy

Department

Civil Engineering

Committee Chair

Stephanie Ivey

Committee Member

Roger Meier

Committee Member

Ricardo Taborda

Committee Member

Sabyasachee Mishra

Abstract

With the rapid advance of information technology in the transportation industry, of which intermodal transportation is one of the most important subfields, the scale and dimension of problem sizes and datasets is rising significantly. This trend raises the need for study on improving the efficiency, profitability and level of competitiveness of intermodal transportation networks while exploiting the rich information of big data related to these networks. Therefore, this dissertation aims to investigate intermodal transportation network design problems, especially practical optimization problems, and to develop more realistic and effective models and solution approaches that will assist network operators and/or decision makers of the intermodal transportation system. This dissertation focuses on developing a novel strategy for solving the Minimum Cost Flow (MCF) problem for large-scale network design problems by adopting a divide-and-conquer policy during the optimization process. The main contribution is the development of an agglomerative clustering based tiling strategy to significantly reduce the computational and peak memory consumption of the MCF model for large-scale networks. The tiling strategy is supported by the regional-division theorem and -approximation regional-division theorem that are proposed and proved in this dissertation. The region-division theorem is a sufficient condition to exactly guarantee the consistency between the local MCF solution of each sub-network obtained by the aforementioned tiling strategy and the global MCF solution of the whole network. Furthermore, the -approximation region-division theorem provides worst-case bounds, so that the practical approximation MCF solution closely approximates the optimal solution in terms of its optimal value. A series of experiments are performed to evaluate the utility of the proposed approach of solving the large-scale MCF problem. The results indicate that the proposed approach is beneficial to save the execution time and peak memory consumption in large-scale MCF problems under different circumstances.

Comments

Data is provided by the student.

Library Comment

Dissertation or thesis originally submitted to the local University of Memphis Electronic Theses & dissertation (ETD) Repository.

Download

Share

COinS