Using advanced tabu search techniques to solve airline disruption management problems

iv Acknowledgements I would like to express my sincerest gratitude to my supervisor, Dr. J. Wesley Barnes, who guided me through my graduate study. His insightful guidance and unwavering support made this dissertation possible. I am also indebted to his continuous encouragement and inspiration. I would like to thank all the other members of my dissertation committee, Dr. T. Kutanoglu, for their invaluable comments and suggestions. My thanks also go to Dr. Melba M. Crawford, for her support and guidance in the early stages of my doctoral study. I would also like to thank Michael Argüello and Benjamin Thengvall for providing information about their earlier research work. I would like to thank my friends for offering me suggestions with regard to doctoral study based on their own experience. They are: I must thank my family for their encouragement and support. My parents-in-law helped with childcare and housework. My husband Wanwan encouraged me throughout my study. I highly appreciate what they have done for me in the past few years. And to my son Kevin, thank you for bringing much joy to me. Disruption Management in the airline industry plays an important role in airline operations. The goal of disruption management is to minimize the costs associated with disruptions while returning to the original schedule. Methodologies using advanced tabu search (TS) were investigated to solve two flight rescheduling problems: the aircraft grounding problem and the reduced station capacity problem. The objectives of both problems were to minimize the schedule recovery costs associated with flight schedule modifications and deviations from the original route, which are composed of the sum of delay costs, cancellation costs and aircraft route swap costs. Reflecting the cost of the deviation from the original route, the swap cost was modeled as a non-linear function of the swaps of aircraft between routes. In each problem, a stand-alone tabu search approach was constructed to holistically minimize the sum of the cost of delays, cancellations and swaps. Next a hybrid method which combined a time-space network flow model with side constraints and a limited tabu search was created which attacked the problem in two steps: first, the total cost of delays and cancellations was minimized by the network flow vi model; second, a limited tabu search was conducted to minimize the number of swaps. A second hybrid method was then developed, which utilized the result from the first hybrid method as …