Node-Reliability

Monte Carlo, Laplace, and Stochastic Approximations and a Greedy Link-Augmentation Strategy

Journal Article (2025)
Author(s)

X. Liu (TU Delft - Network Architectures and Services)

Robert Kooij (TNO, TU Delft - Quantum & Computer Engineering)

P.F.A. Van Mieghem (TU Delft - Network Architectures and Services)

Department
Quantum & Computer Engineering
DOI related publication
https://doi.org/10.1109/TNSM.2025.3607004
More Info
expand_more
Publication Year
2025
Language
English
Department
Quantum & Computer Engineering
Volume number
23
Pages (from-to)
756-766
Reuse Rights

Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.

Abstract

The node-reliability polynomial nRelG(p) measures the probability that a connected network remains connected given that each node functions independently with probability p. Computing node-reliability polynomials nRelG(p) exactly is NP-hard. Here we propose efficient approximations. First, we develop an accurate Monte Carlo simulation, which is accelerated by incorporating a Laplace approximation that captures the polynomial’s main behavior. We also introduce three degree-based stochastic approximations (Laplace, arithmetic, and geometric), which leverage the degree distribution to estimate nRelG(p) with low complexity. Beyond approximations, our framework addresses the reliability-based Global Robustness Improvement Problem (k-GRIP) by selecting exactly k links to add to a given graph so as to maximize its node reliability. A Greedy Lowest-Degree Pairing Link Addition (Greedy-LD) Algorithm, is proposed which offers a computationally efficient and practically effective heuristic, particularly suitable for large-scale networks.

Files

Taverne
warning

File under embargo until 09-03-2026