An agent-based model for modal shift in public transport

TL;DR

This paper presents an agent-based model to analyze how behavioral factors influence modal shift in public transport, revealing complex congestion patterns and optimizing mode choices through genetic algorithms.

Contribution

It introduces a stylized agent-based model to study behavioral impacts on modal shift and congestion, with application to a real Paris train network case study.

Findings

Non-trivial congestion patterns emerge based on behavioral parameters.

Optimal mode choice compromises can be identified using genetic algorithms.

Behavioral parameters significantly influence congestion outcomes.

Abstract

Modal shift in public transport as a consequence of a disruption on a line has in some cases unforeseen consequences such as an increase in congestion in the rest of the network. How information is provided to users and their behavior plays a central role in such configurations. We introduce here a simple and stylised agent-based model aimed at understanding the impact of behavioural parameters on modal shift. The model is applied on a case study based on a stated preference survey for a segment of Paris suburban train network. We systematically explore the parameter space and show non-trivial patterns of congestion for some values of discrete choice parameters linked to perceived wait time and congestion. We also apply a genetic optimisation algorithm to the model to search for optimal compromises between congestion in different modes.