From delay to disruption: the impact of service degradation on public transport network

Abstract

Passengers in public transport networks (PTN) experience every day different kinds of disturbances, from small delays to big service interruptions. A strict boundary between delays and disruptions is not plausible, but rather there is a continuous range of disruptions with different characteristics and impact. Despite this, in disruption analysis the typical definition of disruption is a link closure for a certain amount of time. No attention is given to the relationship between delays and disruptions or to short-term disruptions. This is of particular interest in multimodal PTN, where a link closure is not often observable. The aim of this paper is to link delays to disruptions and identify the relationships between the disruption characteristics and their impact on the users. Real disturbances of the Zürich PTN are analysed to identify disruptions with different characteristics. Therefore, the disruption impact is observed on simulated passengers’ paths. Using machine learning techniques, a strong dependence between the characteristics and the impact is identified and the most important features are highlighted.