Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 41 of the 44 references it cites were found in a registry or an index; 6 of the 14 sources its claims cite were checked for support.

Mitigation of cascading failures in complex networks

Open access

Abstract

AbstractCascading failures in many systems such as infrastructures or financial networks can lead to catastrophic system collapse. We develop here an intuitive, powerful and simple-to-implement approach for mitigation of cascading failures on complex networks based on local network structure. We offer an algorithm to select critical nodes, the protection of which ensures better survival of the network. We demonstrate the strength of our approach compared to various standard mitigation techniques. We show the efficacy of our method on various network structures and failure mechanisms, and finally demonstrate its merit on an example of a real network of financial holdings.

Overview

CoChat Summary

TL;DR

The paper proposes a local-structure mitigation algorithm that protects fragile nodes with fragile neighbors to stop cascading failures in complex networks. It reports improved survival over random, degree-based, and centrality-based strategies across model networks and a European bank–sovereign debt network.

Key findings

  • Protecting nodes selected by the algorithm raises survival above the unprotected critical threshold; for the ER network the unprotected critical threshold is just below 0.86, and partial protection at 70% or 40% still noticeably increases survival probability.
  • The method outperforms random, highest-degree, lowest-degree, and centrality-based protection; for ER networks, surprisingly, low-degree protection outperforms high-degree protection, which under-performs even random selection.
  • Near the critical threshold the protected set is small, of the order of several percent of the network, and the system tends to either collapse completely or survive almost entirely.
  • In the European banks’ sovereign debt exposure network (close to 90 European banks, 2011), protecting around fifteen financial institutions is enough for a 14% impact threshold; for severe failures such as Italy or Spain, holding less than twenty institutions safe spares more than half the network, while at a 6% failure threshold roughly a third of all financial institutions must be protected.

Methods

  • Simulations use undirected, unweighted networks—mainly ER and a bipartite bank–asset structure, with scale-free and interdependent networks in the Supplementary—typically of order 10k nodes, repeated at least 150 times and up to 1,000; the main failure mechanism is fractional threshold, with k-core and interdependent failure also considered.
  • The mitigation algorithm defines a node as fragile under failure mechanism M when removal of a minimal relevant quantity K of its links causes failure; it selects fragile nodes that also have at least some fragile neighbors, then applies a heuristic refinement based on the induced subgraph’s minimal degree.
  • Cascades are initiated by removing a single node while selected nodes are held safe; for the real financial network, a sovereign debt is set to zero and banks fail if remaining holdings fall below a specified threshold relative to initial holdings, with only banks protected.

Limitations

  • Stated: the approach lacks theoretical foundation, and a top-down theory could refine the heuristics and parameter choices; several network configurations and failure mechanisms were tested, but real and theoretical networks have more intricacies, including correlated networks, communities, multiplex or time- and space-dependent financial relations, uniform failure thresholds, short-cycle feedback, and healing or new failures.
  • Stated: the real financial application is a very simplistic view of an intricate ecosystem and requires further analysis.
  • Evident: the real network is much smaller than the simulated networks, limiting discussion of distributions; simulations are unweighted and undirected for tractability even though the real network is weighted; the full spectrum of model parameters is left for further investigation.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1038/s41598-020-72771-4

Cites

41 of 41 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Figure 1 Description of the fractional cascade process, colors illustrate the immunization algorithm described in “Mitigation strategy”. (a) The gray node is impacted and will be removed; (b) when removed, all its edges disappear as well; (c) here the important difference between the blue (low degree), green (high degree) and red (medium degree) nodes is highlighted. The blue node has only one additional neighbor, so its failure does not exacerbate the cascade. The green node is of a high degree, so it is not affected by the single neighbor’s failure. The red nodes, however, are vulnerable with respect to a threshold failure perspective, but also have a large number of neighbors that will be affected by their failure. They are the ones enhancing the cascading process; (d,e) show the termination of the failure on the toy model example with only the high-degree node (green) surviving the process.
Figure 1 Description of the fractional cascade process, colors illustrate the immunization algorithm described in “Mitigation strategy”. (a) The gray node is impacted and will be removed; (b) when removed, all its edges disappear as well; (c) here the important difference between the blue (low degree), green (high degree) and red (medium degree) nodes is highlighted. The blue node has only one additional neighbor, so its failure does not exacerbate the cascade. The green node is of a high degree, so it is not affected by the single neighbor’s failure. The red nodes, however, are vulnerable with respect to a threshold failure perspective, but also have a large number of neighbors that will be affected by their failure. They are the ones enhancing the cascading process; (d,e) show the termination of the failure on the toy model example with only the high-degree node (green) surviving the process.CC BY · PubMed Central
Figure 2 Comparison of the effect of probabilistic protection between several network structures. (a) ER network with 10,000 nodes and ⟨k⟩=8. The blue, purple, orange and yellow dotted-dashed lines represent a probability of 1, 0.7, 0.4 and 0 resp. for a chosen node to be protected. Blue and yellow lines represent full and no protection, respectively. The dashed red line shows the fraction of fully protected nodes with respect to the whole network. (b) Assuming four specific thresholds (vertical lines of matching colors to (a)) shown are the probability of system survival for all protection probabilities from 0 to 1; (c,d) same as (a,b) for the bipartite configuration described in “Network structure”.
Figure 2 Comparison of the effect of probabilistic protection between several network structures. (a) ER network with 10,000 nodes and ⟨k⟩=8. The blue, purple, orange and yellow dotted-dashed lines represent a probability of 1, 0.7, 0.4 and 0 resp. for a chosen node to be protected. Blue and yellow lines represent full and no protection, respectively. The dashed red line shows the fraction of fully protected nodes with respect to the whole network. (b) Assuming four specific thresholds (vertical lines of matching colors to (a)) shown are the probability of system survival for all protection probabilities from 0 to 1; (c,d) same as (a,b) for the bipartite configuration described in “Network structure”.CC BY · PubMed Central
Figure 3 Different immunization approaches. (a) ER network with 10,000 nodes and ⟨k⟩=8. The blue, purple, orange and yellow dotted-dashed lines represent respectively, our strategy, a random selection of nodes, highest degree and lowest degree, (c) A bipartite network as described above under the same conditions as in (a); (b,d) follow the network structure of (a,c) resp. while comparing to centrality-based node selection, with purple, orange and yellow dashed-dotted lines corresponding to Closeness, Katz and Betweenness centrality, respectively.
Figure 3 Different immunization approaches. (a) ER network with 10,000 nodes and ⟨k⟩=8. The blue, purple, orange and yellow dotted-dashed lines represent respectively, our strategy, a random selection of nodes, highest degree and lowest degree, (c) A bipartite network as described above under the same conditions as in (a); (b,d) follow the network structure of (a,c) resp. while comparing to centrality-based node selection, with purple, orange and yellow dashed-dotted lines corresponding to Closeness, Katz and Betweenness centrality, respectively.CC BY · PubMed Central
Figure 4 Financial network structure. The left side are various sovereign debt instruments. The right side are banks with their national belonging indicated by the leading letters. Node size indicates Dollar value; color—betweenness centrality. Edge width indicates investment size.
Figure 4 Financial network structure. The left side are various sovereign debt instruments. The right side are banks with their national belonging indicated by the leading letters. Node size indicates Dollar value; color—betweenness centrality. Edge width indicates investment size.CC BY · PubMed Central
Figure 5 (a) Unprotected network. The horizontal axis shows the different failing debt instruments where, blue, orange and green bars correspond to (1-thresholds) of 14%, 10% and 6% resp. Height of the bar corresponds to the total number of surviving nodes (banks and assets). (b) The same network and failure process, but this time when applying our mitigation strategy showing a much higher survival rate for all initial impacts and thresholds tested. (c) Number of banks selected for protection per fragility threshold. As seen before, the more fragile the network—the more nodes are required for protection.
Figure 5 (a) Unprotected network. The horizontal axis shows the different failing debt instruments where, blue, orange and green bars correspond to (1-thresholds) of 14%, 10% and 6% resp. Height of the bar corresponds to the total number of surviving nodes (banks and assets). (b) The same network and failure process, but this time when applying our mitigation strategy showing a much higher survival rate for all initial impacts and thresholds tested. (c) Number of banks selected for protection per fragility threshold. As seen before, the more fragile the network—the more nodes are required for protection.CC BY · PubMed Central

Verification

Claims the paper makes

2 Valid · 2 Not valid · 4 Not checked · 76 Own results and interpretations

48 more claims

Create a free account to see all 84 claims and every source verdict.

See all claims — free account
  • Probably not a substantial claim. 15
  • It may state two separate claims as one. 19

Create a free account to see these claims.

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

1 of 1

Citations

Cites

Show all 44 references

Cited by

See all 67

Metadata

Authors
Smolyak, A., Levy, O., Vodenska, I., Buldyrev, S., & Havlin, S.
Affiliations
  • Bar-Ilan University (IL)
  • Boston University (US)
  • Yeshiva University (US)
Corresponding author
Smolyak, Alex
Venue
Scientific Reports
Published
2020-09-30
Volume
10
Issue
1
Pages
16124-16124
ISSN
2045-2322 (ISSN-L)
OpenAlex
W3091426098
Cited by
67
References
44
Citations by year
2026: 10 · 2025: 11 · 2024: 18 · 2023: 12 · 2022: 8 · 2021: 6 · 2020: 1
FWCI
2.85
Type
journal article
Language
en
Open access
Licence
Locations
  • Scientific Reports (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • OpenBU (Boston University) (publishedVersion, cc-by)
Funders
  • Israel Science Foundation
  • Defense Threat Reduction Agency (HDTRA-1-19-10016)
  • Center for Research in Applied Cryptography and Cyber Security, Bar-Ilan University
Topics
  • Physical Sciences › Physics and Astronomy › Statistical and Nonlinear Physics › Complex Network Analysis Techniques
  • Physical Sciences › Engineering › Civil and Structural Engineering › Infrastructure Resilience and Vulnerability Analysis
  • Social Sciences › Economics, Econometrics and Finance › Economics and Econometrics › Complex Systems and Time Series Analysis
Keywords
cascading failures, complex networks, failure mitigation, network robustness, local network structure, financial networks, infrastructure networks

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-11, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-11 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Mitigation of cascading failures in complex networks. https://app.cochat.ai/papers/10.1038/s41598-020-72771-4
The paper
Smolyak, A., Levy, O., Vodenska, I., Buldyrev, S., & Havlin, S. (2020). Mitigation of cascading failures in complex networks. Scientific Reports. https://doi.org/10.1038/s41598-020-72771-4

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.