Determining service provider and transport system related effects of ridesourcing services by simulation within the travel demand model mobiTopp

Abstract

Purpose Ridesourcing services have become popular recently and play a crucial role in Mobility as a Service (MaaS) offers. With their increasing importance, the need arises to integrate them into travel demand models to investigate transport system-related effects. As strong interdependencies between different people’s choices exist, microscopic and agent-based model approaches are especially suitable for their simulation. Method This paper presents the integration of shared and non-shared ridesourcing services (i.e., ride-hailing and ride-pooling) into the agent-based travel demand model mobiTopp. We include a simple vehicle allocation and fleet control component and extend the mode choice by the ridesourcing service. Thus, ridesourcing is integrated into the decision-making processes on an agent’s level, based on the system’s specific current performance, considering current waiting times and detours, among other data. Results and Discussion In this paper, we analyze the results concerning provider-related figures such as the number of bookings, trip times, and occupation rates, as well as effects on other travel modes. We performed simulation runs in an exemplary scenario with several variations with up to 1600 vehicles for the city of Stuttgart, Germany. This extension for mobiTopp provides insights into interdependencies between ridesourcing services and other travel modes and may help design and regulate ridesourcing services.

Developing mobility as a service – user, operator and governance perspectives

“…In order to analyze the effects of mobility services on transport systems, it is necessary to incorporate travel chains into travel demand models. Wilkes et al [20] show how ridesourcing services can be simulated in a microscopic travel demand model. They highlight that decision parameters for ridesourcing may differ from those for public transport and more data on willingness to pay and travel times is needed to improve the modelling.…”

“…Secondly, three papers provide pilot and survey results with a focus on rural, urban and work-related mobility [17][18][19]. Thirdly, there are four papers which explore including new mobility offerings, in particular ridesourcing and automated transport, into mobility services [20][21][22][23].…”

Do Ridesourcing Platforms Need Socio-Economic Inequality to Thrive?

“…Previous agent-based modelling exercises have provided insights into ridesourcing as a first / last mile solution for public transport 28,29 , driver earnings 25,28,30 , platform pricing policies 25,28,30,31 , the effect of labour market characteristics 25,28,30 , ride-pooling system performance [32][33][34] , and the design of charging infrastructure for an electric fleet 35,36 . An agent-based representation of the ridesourcing market was also used for studying income inequalities resulting from ridesourcing operations 19 .…”

Implementing a Ride-sharing Algorithm in the German National Transport Model (DEMO)

“…Other important research was conducted with the implementation of on-demand ride-sharing using MATSim for the city of Austin ( 5 ), including economical analyses and fleet sizing. Wilkes et al ( 6 ) used an agent-based travel demand model to analyze the effects of new mobility services like ride-sharing for the city of Stuttgart. The successful application of these algorithms in microscopic simulations gives great insight into the requirements for modeling MaaS; transferring these methods to macroscopic models is, however, not a trivial task.…”