Measuring and evolution analyzing of the reliability in urban public transport composite network

Qingyu Luo, Haiyue Yu, Lili Yang, Tingting GaoPLoS 2026 · DOI

TL;DR

This paper introduces a novel dual-layer reliability framework for urban public transport networks, combining connectivity and transport capacity. Using a Coupled Map Lattice (CML) model, they simulate cascading failures and reveal that transport capacity is more sensitive than connectivity, with a surprising temporary rebound in capacity reliability under moderate disruptions. This work offers crucial insights for improving the resilience and operational efficiency of complex, multimodal transit systems.

What Problem Does This Paper Address?

Urban public transport systems are becoming increasingly complex and interconnected, integrating various modes like buses and trains. This complexity, coupled with growing passenger demand and operational dynamics, makes ensuring network reliability a significant challenge. Traditional reliability assessments often focus solely on network topology (connectivity) and overlook crucial dynamic factors like passenger flow and travel delays. This can lead to an overestimation of a network's true resilience. Furthermore, most studies analyze single transit modes, failing to capture the synergistic effects and potential cascading failures that arise from the interdependencies between different transit systems. This paper aims to bridge this gap by developing a more comprehensive reliability evaluation framework that accounts for both structural connectivity and dynamic transport capacity in multimodal networks.

Methods

The authors propose a novel dual-layer reliability framework consisting of two key components:

  • Connectivity Reliability: This measures the network's ability to maintain inter-node connections. It's calculated as the ratio of effective paths between origin-destination (OD) pairs before and after disruptions.
  • Transport Capacity Reliability: This assesses the network's ability to handle passenger flow and maintain timely operations. It's evaluated using two indicators:
  • Passenger Flow-Weighted Betweenness Centrality: This metric refines traditional betweenness centrality by incorporating actual passenger flow, providing a more accurate representation of a node's importance in dissipating passenger traffic.
  • Standard Deviation of Network Delay Time: This captures temporal efficiency by quantifying deviations from historical delay patterns, serving as an indicator of service reliability.

To simulate cascading failures and their evolution, the researchers employ a Coupled Map Lattice (CML) model. This nonlinear dynamical system framework is extended to model not only node states but also link states, reflecting the tightly coupled nature of public transport networks. The CML model simulates how localized disruptions propagate and interact with dynamic feedback effects, allowing for a more realistic representation of failure mechanisms and recovery potential.

Datasets

The study utilizes data for the Nanguan District in Changchun City, China, as a case study. This includes:

  • Urban Rail Transit Data: Information on Metro Line 1, Metro Line 2, and Light Rail Line 3, including station locations and transfer points.
  • Bus Route Data: Information on 105 bus routes and 77 bus stops.
  • Geospatial Data: Obtained via the Gaode Map API to extract inter-station connectivity, travel times, and distances.
  • Real-time Delay Data: Crawled from Gaode Map for individual stations.
  • Passenger Flow Data: Statistical data from the Changchun Municipal Transportation Bureau on average daily passenger volumes for bus and rail transit.

Key Insights and Ideas

  • Transport Capacity is More Sensitive: The study highlights that transport capacity reliability is significantly more vulnerable to disruptions than connectivity reliability. This means that even if a network remains structurally connected, its ability to move passengers efficiently can degrade much faster.
  • The "Rebound" Phenomenon: A surprising finding is the temporary rebound in transport capacity reliability when network damage is between 30% and 50%. This suggests that removing redundant or congested nodes can, in the short term, improve flow distribution efficiency and mitigate overload on critical nodes.
  • Static vs. Dynamic Analysis: The research underscores the limitations of static topological analysis. Dynamic factors like passenger flow and delay times are crucial for accurately assessing network resilience, especially during cascading failures.
  • Importance of Early Intervention: The findings strongly suggest that early-stage intervention to isolate faulty nodes is critical for suppressing fault propagation and maintaining network efficiency.
  • Hub Nodes are Key: High-betweenness nodes, especially those with significant passenger flow, play a crucial role in network resilience. Prioritizing their recovery or protection is vital.

Key Results

  • Connectivity vs. Capacity Decline: Under severe targeted attacks, the failure of just 20% of key nodes resulted in a 60% drop in transport capacity, while connectivity reliability saw a smaller decline. The degradation rate of transport capacity was 1.8 times that of connectivity.
  • Thresholds for Collapse: When network damage exceeds 50%, both connectivity and transport capacity reliability indicators approach zero, indicating a near-total system collapse.
  • Coupling Strength Impact: Higher coupling strength (ε) between urban rail and bus networks generally leads to greater vulnerability and lower reliability, especially at lower coupling values. However, beyond a certain threshold (ε > 0.60), increased coupling can surprisingly improve reliability.
  • Case Study Findings: In the Changchun case study, high-reliability stations like Weixing Square (0.80) and Jiefang Avenue (0.78) demonstrated strong resilience, often serving as major hubs. Conversely, low-reliability stations like Fanrong Road (0.55) were more susceptible to disruptions.

Limitations and Future Work

  • Generalizability: While the Changchun case study represents a typical multimodal network, the authors acknowledge that it doesn't cover the entire city network, which might affect the generalizability of conclusions. Future work aims to expand to more comprehensive urban road networks.
  • Expanded Evaluation Framework: The authors suggest incorporating additional multidimensional indicators in future work, such as passenger satisfaction and environmental carrying capacity.
  • Advanced Technologies: Integrating digital twin technology and reinforcement learning is proposed for developing real-time early warning systems and enabling resilient simulations under complex scenarios.
  • Dynamic Monitoring: The paper advocates for a shift from static assessment to dynamic monitoring and intelligent control of public transport network reliability.

Summary updated

Abstract

With the expansion of urban public transport systems and the increasing integration of multiple modes, ensuring the reliability of composite public transport network has become a critical challenge. To overcome the limitations of traditional reliability research, which primarily relies on connec-tivity analysis, this paper proposes an innovative indicator system integrating connectivity reliability with a newly constructed transport capacity reliability dimension. The latter incorporates dynamic metrics such as passenger load and time delay to more comprehensively characterize actual operational conditions. By establishing a Coupled Map Lattice (CML) model to simulate the cascading propagation of node and line failures under multimodal coupling effects, this study reveals the failure mechanisms and recovery potential of complex networks under dynamic disturbances. Taking Changchun as a case study, the analysis reveals that both connectivity and transport capacity reliability experience minimal decline in the initial stages of disruption. When the scale of network damage exceeds 30%, transport capacity reliability shows a slight rebound. However, once the damage surpasses 50%, both indicators approach zero. The findings suggest that isolating faulty nodes in the early stages and prioritizing the recovery of high-betweenness nodes in the mid-stage can help maintain network operational efficiency. These results provide important theoretical guidance for enhancing the operational reliability and service quality of public transport systems. By applying complex network theory and data-driven methodologies, this research contributes to improving the resilience and sustainability of urban transport systems.
BibTeX
@inproceedings{luo2026measuring,
  author = {Qingyu Luo and Haiyue Yu and Lili Yang and Tingting Gao},
  title = {{Measuring and evolution analyzing of the reliability in urban public transport composite network}},
  booktitle = {PLoS},
  year = {2026},
  doi = {10.1371/journal.pone.0340590},
}
Figure