D.Com. / When low cost carriers are introduced into domestic or regional scheduled air transport markets, the effects tend to be profound. In most markets where they have been introduced, lower prices have tended to lead to the stimulation of demand. As the scope of the market increases, so too does the number of entrants in the market, resulting not only in higher levels of competition but also lower prices and services. The success of the low cost model is indicated by the uptake in the air transport markets, where low cost carriers sometimes account for as much as 50% of the total air traffic movements. The presence of the low cost carriers is not necessarily a guarantee of success and the market failure rates tend to be high...

In this dissertation, I present three theoretical results with real-world applications related to scheduling and distributionally-robust games, important fields in discrete optimization, and computer science. The first chapter provides simple, technology-free interventions to manage elevator queues in high-rise buildings when passenger demand far exceeds the capacity of the elevator system. The problem was motivated by the need to manage passengers safely in light of reduced elevator capacities during the COVID-19 pandemic. We use mathematical modeling, epidemiological expertise, and simulation to design and evaluate our algorithmic solutions. The key idea is to explicitly or implicitly group passengers that are going to the same floor into the same elevator as much as possible, substantiated theoretically using a technique from queuing theory known as stability analysis. This chapter is joint work with Charles Branas, Adam Elmachtoub, Clifford Stein, and Yeqing Zhou, directly in collaboration with the New York City Mayor’s Office of the Chief Technology Officer and the Department of Citywide Administrative Services. The second chapter proposes new algorithms for recomputing passenger itineraries for airlines during major disruptions when carefully planned schedules are thrown into disarray. An airline network is a massive temporal graph, often with tight regulatory and operational constraints. When disruptions propagate through an airline network, the objective is to \textit{recover} within a given time frame from a disruption, meaning we replan schedules affected by the disruption such that the new schedules have to match the originally planned schedules after the time frame. We aim to solve the large-scale airline recovery problem with quick, user-independent, consistent, and near-optimal algorithms. We provide new algorithms to solve the passenger recovery problem, given recovered flight and crew solutions. We build a preprocessing step and construct an Integer Program as well as a network-based approach based on solving multiple-label shortest path problems. Experiments show the tractability of our proposed algorithms on airline data sets with heavy flight disruptions. This chapter is joint work with Clifford Stein, stemming from an internship and collaboration with the Machine Learning team (Artificial Intelligence organization) of GE Global Research, Niskayuna, New York. The third chapter is about computing distributionally-robust strategies for a popular game theory model called Stackelberg games, where one player, called the leader, is able to commit to a strategy first, assuming the other player(s), called follower(s) would best respond to the strategy. In many of the real-world applications of Stackelberg games, parameters such as payoffs of the follower(s) are not known with certainty. Distributionally-robust optimization allows a distribution over possible model parameters, where this distribution comes from a set of possible distributions. The goal for the leader is to maximize their expected utility with respect to the worst-case distribution from the set. We initiate the study of distributionally-robust models for Stackelberg games, show that a distributionally-robust Stackelberg equilibrium always exists across a wide array of uncertainty models, and provide tractable algorithms for some general settings with experimental results. This chapter is joint work with Christian Kroer.

This study attempted to establish the travel behaviours and choice criteria of the South African domestic air passenger and how they differed between low-cost carriers (LCCs) and full-service carriers (FSCs). The study was quantitative and used structured questionnaires to collect data via personal interviews. Descriptive and inferential techniques were used to analyse the data, including a binomial logistic regression to identify predictors of model choice. Analysis This study attempted to establish the travel behaviours and choice criteria of the South African domestic air passenger and how they differed between low-cost carriers (LCCs) and full-service carriers (FSCs). The study was quantitative and used structured questionnaires to collect data via personal interviews. Descriptive and inferential techniques were used to analyse the data, including a binomial logistic regression to identify predictors of model choice. Analysis showed that passengers had a limited understanding of the functioning of the models. This results in consumer perceptions and expectations being discordant with the true differences. In distinguishing between models, LCC passengers rate LCCs more favourably than FSC passengers, but both rate FSCs higher than LCCs. This shows the need of consumers to have the features and services of the FSCs. Amongst the key findings was the absolute importance of price to the passengers on both models when purchasing the ticket. The analysis showed that LCC passengers are highly price sensitive and show loyalty to the lowest price (not airline model). It was apparent that frequent flyer programmes (FFP), or linkages to 3rd party loyalty programmes, for LCCs need to be reconsidered. Younger LCC passengers especially, indicated a need for a simple FFP to receive some form of ‘reward’, as well as benefits traditionally only offered by FSCs. FSC passengers show a greater degree of loyalty and less fare sensitivity. This provides the FSCs with a degree of fare flexibility and the opportunity to move their loyal, less price-sensitive consumers up the price curve to maximise revenue. It was shown that, in distinguishing themselves from FSCs, it is important that LCCs are perceived as being more affordable than FSCs and are offering a value-for-money service. In essence, LCCs have to defend their positioning by (i) ensuring that their fares are not perceived to be as high as a FSCs and (ii) watching that the FSC fares are not declining to a level where FSCs are perceived as being as cheap as a LCC. For LCCs, brand building strategies around issues other than fare need to be devised, with attention paid to identifying determinant factors. / Business Management / D. Com (Business Management)