Neue Publikation in Expert Systems With Applications

9. Mai 2023

Unsere Studie "Reliable route planning and time savings in real-world urban intermodal transportation networks: Evidence from Hamburg, Germany" wurde zur Publikation beim Journal Expert Systems With Applications angenommen.

Autoren: Matthias Ruß, Gunther Gust

Abstract:

Cities around the world suffer from congestion and pollution caused by road traffic. Intermodal transport, which integrates multiple modes in a single journey, is a promising concept for shifting commuters away from private cars and making urban traffic more sustainable. Urban intermodal transport requires reliable route planners—i. e., planners identifying routes that have a high probability of being fast—because reliability matters for commuters. Designing such reliable route planners for the comprehensive set of modes in modern, real-world urban transportation networks is complex. Challenges arise due to the uncertainties inherent in travel times, transfer connections, availabilities of shared vehicles, and the location of shared vehicles. In this paper, we overcome these challenges by providing a reliable route planner that includes the comprehensive set of modes present in real-world urban transportation networks, including cars, trains and buses, station-based bikesharing, free-floating shared e-scooters, as well as walking. As a second contribution, we provide an in-depth analysis of the potential travel-time savings that such a reliable planner generates. Thereby, we generate new knowledge regarding the effect of several parameters and route characteristics on travel time savings—such as the required level of on-time arrival probability, traffic congestion, trip complexity, e-scooter availability, route costs, and more. Our findings are generated based on a large-scale empirical case study in the city of Hamburg covering more than 320,000 travel connections between 730 locations. Reliable planning yields in our base scenario average travel time savings of 8.9% in comparison to a conventional planner. For single routes, these savings reach up to 37.1%. We furthermore find that time savings increase with the required level of on-time arrival probability and the level of traffic. Moreover, the savings increase with the number of rides included in a trip; however, trip distance has no clear effect. Constrained cost budgets of commuters can considerably prolong travel times, whereas the effect of e-scooter availability is comparatively small. For transportation research, these findings extend current knowledge on travel time savings enabled by reliable planning. For transportation practice, our methodology can provide the basis for implementations in routing information systems, such as Google Maps, as well as planning decision support systems for transportation authorities to evaluate interventions, such as investing in the reliability of public transit or adding buffer times to schedules.