Abstract

The core task in passengertransportation systems planning can be described as one of matching thenetwork-wide demand with adequate capacity by deploying resources over thecorresponding parts of the network. A prime challenge inherent in this task isthe uncertainty in demand and supply in transportation systems. Demanduncertainty manifests itself in the form of the passenger choice behavior.Supply uncertainty presents itself as delays and disruptions in resourceavailability. This thesis tackles a set of three related research questionswithin passenger transportation systems while explicitly incorporating theinherent uncertainty in both demand and supply during the planning phases ofair and urban transportation systems.

Flight delay propagationresults in enormous additional operating costs for the airlines, passengers andthe aviation system as a whole. The first part of this thesis is focused onproposing, optimizing and validating a methodological framework for estimatingthe extent of crew-propagated delays and disruptions. We identify the factors thatinfluence the extent of crew-propagated delays, and incorporate them into arobust crew scheduling model. We then develop a fast heuristic approach forsolving the inverse of this robust crew scheduling problem to generate crewschedules that are similar to real-world crew scheduling samples. Along withvarious other findings, our results show that airlines avoid up to 80% ofcrew-related delays through advanced planning methods.

The second part of thisthesis introduces an original integrated optimization approach to comprehensiveflight timetabling and fleet assignment under endogenous passenger choice. Theresulting optimization model is formulated as a mixed-integer linear program.We propose an original multi-phase solution approach, which effectivelycombines several heuristics, to optimize the network-wide timetable of a majorairline within a realistic computational budget. Using case study data fromAlaska Airlines, computational results suggest that the combination of thismodel formulation and solution approaches can result in significant profitimprovements, as compared to the most advanced incremental approaches to flighttimetabling. Additional computational experiments based on several extensionsalso demonstrate the benefits of this modeling and computational framework tosupport various types of strategic airline decision-making in the context offrequency planning, revenue management, and post-merger integration.

The popularity ofridehailing services like Didi, Lyft, and Uber has soared recently. Whileridehailing services offer a very convenient mode of urban transportation formany passengers, these additional vehicles on the road contribute to urban roadtraffic congestion, and in some cases are held responsible for falling public transitridership. A high quality public transit system is an effective means toalleviate urban road traffic congestion, but the interdependence betweentransit ridership, ridehailing ridership and urban road traffic congestionmotivates the following question: can public transit and ridehailing co-existand thrive to improve the overall social welfare? To answer this question, wedevelop a novel mixed-integer nonlinear optimization model, and a new set ofsolution algorithms to optimize transit timetables. Our model explicitlyaccounts for the impacts on road congestion and the passengers' mode-switchingbehavior between transit and ridehailing services.

Time

12月25日（星期五）10:00-11:30

Speaker

Keji Wei received PhD fromthe operations research at Dartmouth at 2019. His current research interest islies at the intersection of transportation and optimization, particularly inaviation area. He is the recipient of the AGIFORS Anna Valicek Award in 2017and 2019. He is also the winner of Best Dissertation Award in aviationassociation section in Informs 2020. Before coming to Dartmouth, he obtainedthe Bachelor of Science in Automation from Xian Jiaotong University.

Currently, Keji is a SeniorOperations Research Analyst at Sabre, designing and implementing the operationsresearch techniques for passenger recovery module to serve more than +20airlines in the world.

Venue

Zoom会议室ID:96489843147

密码：583515