Scenarios of Destruction for Large Network and Increasing Reliability

TL;DR

This paper models damage scenarios in large networks using adjacency matrices and Markov chains, and proposes adding high-dimensional random agents to significantly enhance overall network reliability.

Contribution

It introduces a novel approach to network damage modeling with matrix powers and Markov chains, and proposes a reliability augmentation method using high-dimensional random agents.

Findings

Cascade damage scenarios modeled with adjacency matrix powers

Probabilistic damage variants analyzed via Markov chains

Reliability of reinforced networks significantly improved

Abstract

Damage scenarios for large networks are considered. The cascade scenario is described by means of powers of adjacency matrix. More difficult probabilistic variants of the large network damage are modeling by Markov chains. For reliability augmentation of networks we add a set of random intermediate agents with big dimensionality. It provides high reliability of all system even with low reliability of single components. Probabilistic estimation of reliability for reinforced network is made.