Resilience Assessment for Complex Networks Based on Recovery Strategies

Abstract

The vulnerability of complex networks to unexpected disruptive events could be reduced by increasing network resilience through the efficient recovery of the damaged network. To find the most efficient recovery strategy among the existing variety of strategies, a resilience-based framework was proposed and implemented for both localized attacks and cascading failures. For localized attacks, preferential recovery based on nodal weights (PRNW), periphery recovery (PR) and localized recovery (LR) were assessed. Additionally, probability-based recovery (RS1) and recovery of neighboring or boundary nodes (RS2) methods were evaluated for cascading failures. Considering the advantages and disadvantages of these strategies, a hybrid recovery strategy was proposed to achieve high network resilience in a timely manner with a manageable amount of cost. Overall, this study aids in the assessment and the development of a cost-effective resilience-based recovery strategy.