Choice-based Airline Schedule Design And Fleet Assignment Model

Airline scheduling which determine the appropriate flight frequency and timetable ina market, and assign available aircraft to flights in order to match aircraft seating capacityclosely with flight demand, is one of the most important tasks in the Aviation OperationsResearch. But most of the existing airline scheduling model ignores or fails to considerthe impact of airline schedule on passenger demand, which lead to flight does not match tothe actual demand and result in a waste of transport capacity. There are mainly two typesgreater impacts. The first one is the flight frequency can afect the unconstrained demandin a market. And the second one is the passenger demand on a given flight can spill toother flights if it is larger than its flight capacity. We modeled these two types of impactsinto one passenger choice behavior model and integrated it with airline scheduling modelas a unified model, in order to consider the correct interactions between airline scheduleand passenger demand, and better utilize airline’s transport capacity.In this paper, we employed the Multinomial Logit(MNL) Model to model passengerchoice behavior, and took into account the flight frequency, flight fare, departure timeand journey time as passenger choice behavior factors. Then we builded Choice-basedAirline Schedule Design and Fleet Assignment Model(CB-ASDFAM) which integrateMNL Model with Schedule Design Model and Fleet Assignment Model. CB-ASDFAMwhich involve nonlinear constraints and integer variables is a Mixed Integer NonlinearProgramming(MINLP) Model. To achieve fast solution of large-scale real problems, wedesigned and implemented a solution algorithm based on the Branch-and-Bound methodand the Sequential Quadratic Programming method. For the flights network of a domesticmajor airline, our algorithm can be completed in a short time for there airline scheduling.The main contribution of this paper is constructing passenger choice behavior andairline scheduling in one integrated optimization model. Using discrete choice analysis,we accurately depicts the passenger choice behavior and the interaction between airlineschedule and passenger demand. This will provide meaningful guidance for aviationoperation management. Another innovation of this paper is employing the passengerchoice behavior model to model two types impacts of airline scheduling on passengerdemand. Unlike the existing literatures which are modeling these two types impactsisolated, we provided a unified model based on discrete choice analysis which has wide range of application value and academic value.