Author

Sanya Yimsiri

Graduation Semester and Year

2009

Language

English

Document Type

Dissertation

Degree Name

Doctor of Philosophy in Industrial Engineering

Department

Industrial and Manufacturing Systems Engineering

First Advisor

Jamie Rogers

Abstract

Reverse logistics (RL) involves management of activities that include collection, sort/storage, transportation, recovery, disposal and re-distribution. The product return process is more complicated than forward logistics due to presence of multiple reverse distribution channels, individualized returns with small quantities, extended order cycles associated with product exchanges and a variety of recovery and disposition options. Reverse logistics has been gaining interest from many sectors due to rising costs, environmental concern and tougher regulations. As a result, good reverse logistics network design can help business save costs and meet their bottom lines in this competitive global environment. Most of the previous research in reverse logistics network design has focused on minimizing total costs. However, focusing on minimizing total costs alone is not adequate as there are other environmental and economical factors that contribute to increasing total costs. Therefore, in real life, such design problems usually have multiple objectives to be satisfied. In this research, not only total costs but also total transportation are minimized using non-conventional optimization algorithms. Evolutionary Computation (EC) has been gaining attention due to its effectiveness and robustness in searching for a set of non-inferior solutions for a multi-objective problem. Genetic Algorithms (GA) are considered the most well known class of EC. They are stochastic search techniques that mimic the natural evolution process, and do not require prior domain knowledge.The purpose of this research has been to develop a technique to solve the reverse logistics network design problem with multiple objectives approach. The Pareto based Multi-objective Genetic Algorithm (MOGA) was used to obtain non-dominated solutions. A case study was conducted and sensitivity analysis was performed to compare robustness and stability between typical aggregation based multi-objective genetic algorithms and the Pareto based genetic algorithms developed in this research. The outcome shows that the Pareto based genetic algorithms technique provides an efficient design tool for the reverse logistics network design problem with multiple objectives.

Disciplines

Engineering | Operations Research, Systems Engineering and Industrial Engineering

Comments

Degree granted by The University of Texas at Arlington

Share

COinS