A Unified Approach to Evaluation and Routing in Public Transport Systems

TL;DR

This paper introduces a unified optimization framework for evaluating service quality and modeling passenger route choices in public transport, validated with real-world data from Dutch and Swiss railways.

Contribution

It presents a formal, integrated approach that links evaluation and routing, enabling consistent analysis and improved planning of public transport systems.

Findings

Evaluation and routing are interconnected through optimization.

The framework improves line planning and timetable analysis.

Using appropriate models at different planning stages is crucial.

Abstract

Both evaluating the service quality of a public transport system and understanding how passengers choose between modes or routes is imperative for public transport operators, providers of competing mobility services and policy makers. However, the literature does not offer consensus on how either of these tasks should be performed, which can lead to inconsistent or counter-intuitive results. This paper provides a formal treatment on how fundamental elements of public transport systems (route sets, timetables and line plans) can be evaluated consistently, and how travelers distribute over routes. Our main insight is that evaluation and routing are two sides of the same coin: by solving an appropriate optimization model one obtains both the quality of the route set, timetable or line plan (the optimal objective value), and the distribution of the travelers over the routes (the optimal solution itself). The practical relevance of the new framework is demonstrated with several applications that are validated with real data from the Dutch and Swiss railway networks. The measures and route choice models developed in this paper enable planners to create better line plans and to effectively analyze timetables for inefficiencies. The framework also reveals: (i) the importance of using the right model for the right stage of planning, (ii) that it is not always necessary for public transport planners to accurately model travel behavior, especially for high-level planning, and (iii) that combining models in an inconsistent way can have significant negative consequences that are avoided with the new framework.