Source-linked AI summary
A multilayer perspective for the analysis of urban transportation systems
Alberto Aleta, Sandro Meloni, Yamir Moreno
TL;DR
Existing urban transportation studies often simplify multimodal systems into aggregated networks, limiting representation of transfers and interactions between modes. This paper models entire cities as multiplex networks using complementary line-level and mode-level views, analyzes nine European networks, and adds schedule and waiting-time data for a realistic case study. It finds that both views are useful: mode-level layers support analysis of interdependency and resilience, while line-level layers better represent human mobility and transfer timing.
Problem
Existing studies often analyze one transportation mode or merge multiple modes into an aggregated network, losing information about transfers and multimodal interactions.
Method
The paper represents urban transportation systems as multiplex networks using either each line or each transportation mode as a layer, with detailed schedules and transfer times for one city.
Results
Both per-line and per-mode representations are useful and complementary for analyzing transportation-system properties, functioning, and vulnerabilities across 9 urban networks.
Takeaways & Limitations
Mode-level layers are fundamental for studying interdependency and resilience, whereas line-level layers provide a more realistic model of human mobility.
Takeaways & Limitations
The mobility scenarios assume free flow and omit vehicle carrying capacity because passenger-flow data are unavailable.
Abstract
from arXiv · showhide
Public urban mobility systems are composed by several transportation modes connected together. Most studies in urban mobility and planning often ignore the multi-layer nature of transportation systems considering only aggregated versions of this complex scenario. In this work we present a model for the representation of the transportation system of an entire city as a multiplex network. Using two different perspectives, one in which each line is a layer and one in which lines of the same transportation mode are grouped together, we study the interconnected structure of 9 different cities in Europe raging from small towns to mega-cities like London and Berlin highlighting their vulnerabilities and possible improvements. Finally, for the city of Zaragoza in Spain, we also consider data about service schedule and waiting times, which allow us to create a simple yet realistic model for urban mobility able to reproduce real-world facts and to test for network improvements.
Introduction
Urban transportation studies often collapse multiple modes into one network or aggregate lines by mode, losing information about transfers, waiting, and modal importance. The paper proposes using complementary multiplex representations to preserve these different kinds of information across nine urban networks and a detailed city case study.
- Introduction: Aggregating all transportation modes into one network loses information about how modes interact and how transfers occur.
- Introduction: Mode-based layers provide a compact representation for studying resilience and coupling but neglect transfers and waiting times between lines of the same mode.
- Introduction: Line-based layers preserve transfers and synchronization between lines but cannot quantify the importance of each transportation mode.
- Introduction: The proposed model uses both line-level and mode-level multiplex representations to extract different information from urban transportation systems.
- Introduction: The model is tested on 9 urban transportation networks ranging from small cities to megacities, with detailed schedules and transfer times added for one medium-sized city.
Methods
The method represents each transport line as a network layer, with stops as nodes and geographically weighted links between stops served by the same line.
- Each line of each transport mode is modeled as a separate network layer.
- Stops are represented as nodes, and links connect pairs of stops served by the same line.
- The weight of each link is the geographical distance between its connected stops.
Results
Across cities, multiplex structure reveals similar overlap patterns, distinct mode contributions, and trade-offs between mobility importance and disruption resilience. The Zaragoza case study shows how schedule-aware modeling reproduces transfer and walking patterns while testing disruptions and proposed improvements.
- Network structure: Across cities, most nodes occur in one layer (o_i = 2), fewer in two layers (o_i = 4), and only a few in three or more.Despite differences in city size and layer count, overlapping-degree distributions are similar, with physical constraints and citizens’ interests limiting practical overlap.
- Network structure: Nodes with high overlapping degree do not necessarily have the highest activity across transport superlayers, so importance depends on the criterion used.A highly connected bus stop may be easier to relocate during disruption than a metro or tram stop, making structural importance and disruption importance diverge.
- Mode contributions: More than 70% of Madrid’s shortest paths use the metro superlayer, while 20% of Zaragoza’s use the tram despite its single line and 50 nodes.The analysis first evaluates layer interdependency and then adjusts it by each superlayer’s node fraction to account for unequal layer sizes.
- Mode contributions: After size adjustment, tram and metro are disproportionately important because straighter routes connect distant locations, although the result is purely topological.The findings also indicate multimodal mobility: buses cover much of the city while rail modes connect distant or peripheral locations.
- Zaragoza dynamics: Only 35% of simulated individuals reached their destination without transfers, while walking accounted for about one third of total travel time under the Zaragoza scenario.The model used 1,000 individuals per minute, random origins and destinations, and a 5 km/h walking speed between 08:30 and 10:00.
- Zaragoza dynamics: The bus network is more resilient to disruptions, whereas tram lines speed up the complete network and a proposed east–west tram line would benefit western Zaragoza most.Removing the tram increased travel time more than removing the two most-used bus lines, while the proposed addition produced the largest decrease in the city’s west.
Discussion
The paper finds that per-line and per-mode multiplex representations are complementary for analyzing transportation structure, functioning, and vulnerabilities. A schedule- and waiting-time-based model also reproduces real-world facts and supports exploring disruptions and network improvements.
- The study analyzes 9 urban transportation networks and identifies universal properties related to underlying city structure.
- Per-line and per-transportation-mode representations provide complementary information about system functioning and vulnerabilities.
- Detailed service schedules and waiting times produce a realistic urban-mobility model that reproduces real-world facts.
- The model can explore service disruptions and possible network improvements using information generally publicly available for major cities.
Author contributions statement
The passage set records the authors’ roles and financial-interest disclosure, alongside captions documenting the study’s network and mobility visualizations.
- A. A. performed the analysis and numerical simulations, while A. A., S. M., and Y. M. designed the study and analyzed the results.
- The authors state that they wrote, reviewed, and approved the final manuscript.
- The authors declare no competing financial interests.
- The figure captions cover multiplex-network structure, superlayer activity and interdependency, mobility outcomes, and disruption or improvement scenarios.