Skip to main content

Simulation and Evaluation of Urban Rail On-board Crowding

Time: Wed 2019-11-13 10.00

Location: B24, Brinellvägen 23, KTH, Stockholm (English)

Subject area: Transport Systems

Doctoral student: Soumela Peftitsi , Transportplanering

Opponent: Docent Andreas Tapani, VTI

Supervisor: Docent Erik Jenelius, Transportplanering; PhD Oded Cats, Transportplanering

Export to calendar


As travel demand grows in many metropolitan areas, overcrowding in public transport systems has become a major issue. Crowding distribution can be highly uneven over station platforms and multi-car transit vehicles. This implies that cars are not equally utilized and thus, passengers perceive on-board crowding higher. Furthermore, there are higher fleet requirements to serve the demand, leading to higher operating costs. Public transport operators' main concern is to reduce on-board crowding effects through investments. Thus, there is a need to thoroughly understand how passengers are distributed among individual cars and evaluate on-board crowding in a more realistic way, by considering that passenger distribution over vehicles is not even. 

Paper I explores the utilization of automated public transport data to analyze passengers' boarding behavior and investigate the trade-offs between crowding and walking when passengers make car boarding choices. The framework proposed in this paper is applied in a case study of a line segment of Stockholm metro network. In crowded conditions, passengers are found to make trade-offs, increase walking to avoid in-vehicle crowding, when choosing a train car. 

Paper II focuses on the extension of an existing public transit operations and assignment model, BusMezzo, to allow modelling of crowding in individual cars of a multi-car vehicle. The simulation framework is applied in a case study in Stockholm metro network, evaluating the effects of variations in demand and station layout on the distribution of passengers over vehicles. Increased demand results in a more even passenger distribution at the most crowded stations. This is due to less remaining capacity that leads to low flexibility of how passengers are distributed among train cars. The closure of a popular access point at a selected station leads to more evenly loaded trains and lower perceived in-vehicle travel times per passenger.