City-wide shared taxis: A simulation study in Berlin

Bischoff, Joschka; Maciejewski, Michal; Nagel, Kai · 2017 · OpenAlex-citations

DOI: 10.1109/itsc.2017.8317926

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This paper addresses the potential of dynamic ridesharing services to transform urban mobility, specifically investigating whether shared taxi fleets can reduce vehicle kilometers traveled (VKT) while maintaining acceptable service levels. Motivated by the rise of app-based ridesharing and the anticipated future of shared autonomous vehicles, the authors aim to quantify the efficiency gains of pooling passengers in a real-world context. The study focuses on Berlin, Germany, to evaluate how shared taxi operations impact fleet utilization, passenger travel times, and network congestion compared to traditional non-shared taxi services. The researchers employed MATSim, an open-source multi-agent transport simulation framework, to model city-wide taxi demand. They utilized a real-world dataset of taxi requests from Berlin recorded on a typical workday in 2013. The simulation implemented an insertion-based heuristic algorithm for dynamic routing, which assigns new passenger requests to existing vehicle routes by minimizing the increase in total vehicle workload. The model accounted for constraints such as maximum waiting times, travel time detours, and vehicle capacity. The authors conducted 180 simulation runs varying vehicle capacities (2, 3, or 4 passengers) and parameters defining acceptable detour multipliers and fixed time allowances. Performance was evaluated against a non-shared base case using metrics including average wait time, ride time, rejection rate, and the ratio of driven distance to direct trip distance. The results indicate that shared taxi services can reduce overall VKT by 15–22% compared to non-shared operations, while keeping average passenger wait times lower than or comparable to the base case. Travel time increases were minimal, with average detours remaining low. The study found that vehicle capacity of 3 or 4 did not significantly outperform capacity 2, as vehicles rarely carried more than two simultaneous requests due to tight time constraints. Consequently, standard-sized cars were deemed sufficient for most shared operations. Spatially, sharing potential was highest in the city center and areas near Tegel Airport, where demand density supported efficient pooling. The rejection rate for requests was generally lower in shared scenarios, suggesting improved service availability. The significance of this work lies in demonstrating that dynamic ridesharing can substantially improve fleet efficiency and reduce traffic impact without severely compromising passenger convenience. The findings suggest that operators should prioritize shared services in high-demand urban cores and that increasing vehicle size beyond standard cars offers diminishing returns. The study provides a validated simulation framework and empirical evidence supporting the viability of shared taxi models as a step toward sustainable, on-demand mass transport systems.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• exposure measurement