Date of Award

August 2014

Degree Type

Dissertation

Degree Name

Doctor of Philosophy

Department

Management Science

First Advisor

Anthony D. Ross

Committee Members

Anthony D. Ross, Xiang Fang, Sanjoy Ghose, Kaan Kuzu, Matthew Petering

Abstract

Steel manufacturing is both energy and capital intensive, and it includes multiple production stages, such as iron-making, steelmaking, and rolling. This dissertation investigates the order schedule coordination problem in a multi-stage manufacturing context. A mixed-integer linear programming model is proposed to generate operational (up to the minute) schedules for the steelmaking and rolling stages simultaneously. The proposed multi-stage scheduling model in integrated steel manufacturing can provide a broader view of the cost impact on the individual stages. It also extends the current order scheduling literature in steel manufacturing from a single-stage focus to the coordinated multi-stage focus. Experiments are introduced to study the impact of problem size (number of order batches), order due time and demand pattern on solution performance. Preliminary results from small data instances are reported. A novel heuristic algorithm, Wind Driven Algorithm (WDO), is explained in detail, and numerical parameter study is presented. Another well-known and effective heuristic approach based on Particle Swarm Optimization (PSO) is used as a benchmark for performance comparison. Both algorithms are implemented to solve the scheduling model. Results show that WDO outperforms PSO for the proposed model on solving large sample data instances. Novel contributions and future research areas are highlighted in the conclusion.

Share

COinS