Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Predicting Cascade Failures in Interdependent Urban Infrastructure Networks

Yinzhou Tang, Jinghua Piao, Huandong Wang, Shaw Rajib, Yong Li

TL;DR

Cascading failures in interdependent urban infrastructure threaten city function, and existing single-network models fail to capture cross-network coupling. The authors introduce $I^3$, a dual-graph autoencoder framework with three pre-training embeddings (link-prediction, global-pooling, and initial-node enhancement) and a Relational GCN decoder to predict final failed nodes in a heterogeneous interdependent network. The method demonstrates state-of-the-art performance on synthetic and real-world datasets, with substantial gains in AUC, Precision, Recall, F1, and RMSE for cascade volume, and it validates phase-transition behavior and robustness across cascade configurations. The work provides practical implications for resilience planning and real-time monitoring, with code released for reproducibility and further development.

Abstract

Cascading failures (CF) entail component breakdowns spreading through infrastructure networks, causing system-wide collapse. Predicting CFs is of great importance for infrastructure stability and urban function. Despite extensive research on CFs in single networks such as electricity and road networks, interdependencies among diverse infrastructures remain overlooked, and capturing intra-infrastructure CF dynamics amid complex evolutions poses challenges. To address these gaps, we introduce the \textbf{I}ntegrated \textbf{I}nterdependent \textbf{I}nfrastructure CF model ($I^3$), designed to capture CF dynamics both within and across infrastructures. $I^3$ employs a dual GAE with global pooling for intra-infrastructure dynamics and a heterogeneous graph for inter-infrastructure interactions. An initial node enhancement pre-training strategy mitigates GCN-induced over-smoothing. Experiments demonstrate $I^3$ achieves a 31.94\% in terms of AUC, 18.03\% in terms of Precision, 29.17\% in terms of Recall, 22.73\% in terms of F1-score boost in predicting infrastructure failures, and a 28.52\% reduction in terms of RMSE for cascade volume forecasts compared to leading models. It accurately pinpoints phase transitions in interconnected and singular networks, rectifying biases in models tailored for singular networks. Access the code at https://github.com/tsinghua-fib-lab/Icube.

Predicting Cascade Failures in Interdependent Urban Infrastructure Networks

TL;DR

Cascading failures in interdependent urban infrastructure threaten city function, and existing single-network models fail to capture cross-network coupling. The authors introduce , a dual-graph autoencoder framework with three pre-training embeddings (link-prediction, global-pooling, and initial-node enhancement) and a Relational GCN decoder to predict final failed nodes in a heterogeneous interdependent network. The method demonstrates state-of-the-art performance on synthetic and real-world datasets, with substantial gains in AUC, Precision, Recall, F1, and RMSE for cascade volume, and it validates phase-transition behavior and robustness across cascade configurations. The work provides practical implications for resilience planning and real-time monitoring, with code released for reproducibility and further development.

Abstract

Cascading failures (CF) entail component breakdowns spreading through infrastructure networks, causing system-wide collapse. Predicting CFs is of great importance for infrastructure stability and urban function. Despite extensive research on CFs in single networks such as electricity and road networks, interdependencies among diverse infrastructures remain overlooked, and capturing intra-infrastructure CF dynamics amid complex evolutions poses challenges. To address these gaps, we introduce the \textbf{I}ntegrated \textbf{I}nterdependent \textbf{I}nfrastructure CF model (), designed to capture CF dynamics both within and across infrastructures. employs a dual GAE with global pooling for intra-infrastructure dynamics and a heterogeneous graph for inter-infrastructure interactions. An initial node enhancement pre-training strategy mitigates GCN-induced over-smoothing. Experiments demonstrate achieves a 31.94\% in terms of AUC, 18.03\% in terms of Precision, 29.17\% in terms of Recall, 22.73\% in terms of F1-score boost in predicting infrastructure failures, and a 28.52\% reduction in terms of RMSE for cascade volume forecasts compared to leading models. It accurately pinpoints phase transitions in interconnected and singular networks, rectifying biases in models tailored for singular networks. Access the code at https://github.com/tsinghua-fib-lab/Icube.

Paper Structure

This paper contains 26 sections, 27 equations, 5 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Related Works
  3. CF prediction on single network
  4. CF prediction on interdependent network
  5. Problem Formulation
  6. Method
  7. Multi-Task Pre-trained Encode Module
  8. Link Prediction Pre-training
  9. Global Pooling Pre-training
  10. Initial Node Enhancement Pre-training
  11. Graph Feature Encoding
  12. RGCN-Enhanced Decode Module
  13. Experiments
  14. Datasets
  15. Synthetic Dataset
  16. ...and 11 more sections

Figures (5)

  • Figure 1: Illustration of interdependent urban infrastructure networks.
  • Figure 2: Model architecture of $I^3$.
  • Figure 3: Difference between predicted heat map of different models and the ground truth.
  • Figure 4: Experiment of Phase Transition
  • Figure 5: Experiments of Different Single Networks