Failure-recovery model with competition between failures in complex networks: a dynamical approach

TL;DR

This paper introduces a dynamical model for failure and recovery in complex networks, revealing oscillations, sharp decreases, and hysteresis in active node density, linked to failure competition and k-core percolation.

Contribution

It provides a dynamical analysis of failure-recovery models, uncovering oscillatory behavior and phase transitions not previously characterized.

Findings

Systems can exhibit sharp temporal decreases in active nodes.

Oscillatory regimes depend on parameter values and failure competition.

Discontinuous drops in active nodes relate to k-core percolation.

Abstract

Real systems are usually composed by units or nodes whose activity can be interrupted and restored intermittently due to complex interactions not only with the environment, but also with the same system. Majdand\v{z}i\'c $et\;al.$ [Nature Physics 10, 34 (2014)] proposed a model to study systems in which active nodes fail and recover spontaneously in a complex network and found that in the steady state the density of active nodes can exhibit an abrupt transition and hysteresis depending on the values of the parameters. Here we investigate a model of recovery-failure from a dynamical point of view. Using an effective degree approach we find that the systems can exhibit a temporal sharp decrease in the fraction of active nodes. Moreover we show that, depending on the values of the parameters, the fraction of active nodes has an oscillatory regime which we explain as a competition between different failure processes. We also find that in the non-oscillatory regime, the critical fraction of active nodes presents a discontinuous drop which can be related to a "targeted" k-core percolation process. Finally, using mean field equations we analyze the space of parameters at which hysteresis and oscillatory regimes can be found.