A Railway Seat Allocation Model Considering Fairness and Dependent Demand between Consecutive Trains

Abstract

This study propose a railway seat allocation model which considers passengers may choose next train rather than choose another mode immediately after booking rejected This model also considers fairness which means avoiding to seats all allocated to specific ODs and causing some ODs cannot get any resource In order to solve the multi-objective problem this study introduces a genetic algorithm NSGA-II This study also defines demand satisfaction ratio to measure fairness After empirical study we find that considering passengers’ dependent demand really improve expected total revenue and seats are allocated more to high revenue-contributed ODs Then while we consider dependent demand and fairness simultaneously fairness indeed can be improved by reducing limited revenue and low revenue-contributed ODs can get more seats than before This model can provide some information to Railway Company to calculate number of subsidy while local government asks for stopping or allocating more seats to their stations

Date of Award	2016 Nov 14
Original language	English
Supervisor	Yung-Hsiang Cheng (Supervisor)