Date of Award

August 2018

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Industrial Engineering

Committee Member

Sandra Eksioglu

Committee Member

Burak Eksioglu

Committee Member

David Neyens

Committee Member

Kevin Taaffe

Committee Member

Khoa Truong

Abstract

For many years, researchers have focused on developing optimization models to design and manage supply chains. These models have helped companies in different industries to minimize costs, maximize performance while balancing their social and environmental impacts. There is an increasing interest in developing models which optimize supply chain decisions of perishable products. This is mainly because many of the products we use today are perishable, managing their inventory is challenging due to their short shelf life, and out-dated products become waste. Therefore, these supply chain decisions impact profitability and sustainability of companies and the quality of the environment. Perishable products wastage is inevitable when demand is not known beforehand. A number of models in the literature use simulation and probabilistic models to capture supply chain uncertainties. However, when demand distribution cannot be described using standard distributions, probabilistic models are not effective. In this case, using stochastic optimization methods is preferred over obtaining approximate inventory management policies through simulation.

This dissertation proposes models to help businesses and non-prot organizations make inventory replenishment, pricing and transportation decisions that improve the performance of their system. These models focus on perishable products which either deteriorate over time or have a fixed shelf life. The demand and/or supply for these products and/or, the remaining shelf life are stochastic. Stochastic optimization models, including a two-stage stochastic mixed integer linear program, a two-stage stochastic mixed integer non linear program, and a chance constraint program are proposed to capture uncertainties. The objective is to minimize the total replenishment costs which impact prots and service rate. These models are motivated by applications in the vaccine distribution supply chain, and other supply chains used to distribute perishable products.

This dissertation also focuses on developing solution algorithms to solve the proposed optimization models. The computational complexity of these models motivated the development of extensions to standard models used to solve stochastic optimization problems. These algorithms use sample average approximation (SAA) to represent uncertainty. The algorithms proposed are extensions of the stochastic Benders decomposition algorithm, the L-shaped method (LS). These extensions use Gomory mixed integer cuts, mixed-integer rounding cuts, and piecewise linear relaxation of bilinear terms. These extensions lead to the development of linear approximations of the models developed. Computational results reveal that the solution approach presented here outperforms the

standard LS method.

Finally, this dissertation develops case studies using real-life data from the Demographic Health Surveys in Niger and Bangladesh to build predictive models to meet requirements for various childhood immunization vaccines. The results of this study provide support tools for policymakers to design vaccine distribution networks.

Share

COinS
 
 

To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.