Resilience of and recovery strategies for weighted networks

Abstract

The robustness and resilience of complex networks have been widely studied and discussed in both research and industry because today, the diversity of system components and the complexity of the connection between units are increasingly influencing the reliability of complex systems. Previous studies have focused on node failure in networks, proposing several performance indicators. However, a single performance indicator cannot comprehensively measure all the performance aspects; thus, the selected performance indicators and recovery strategies lack consistency with respect to the targeted complex systems. This paper introduces a novel stress–strength-balanced weighted network model based on two network transmission hypotheses. Then, with respect to different concerns of the complex network, we propose two modified network performance measurement indicators and compare these indicators in terms of their trends in the attack process. In addition, we introduce several network recovery strategies and compare their efficiencies. Our findings are as follows: (1) The evaluation and judgment of the network performance depend on the performance measurement indicators we use. (2) Different recovery strategies exhibit distinct efficiencies in recovering different aspects of network performance, and no strategy exists that can improve all the network performance aspects simultaneously. (3) The timing of the recovery is proved to have a deep influence on the cost and efficiency of network recovery; thus, the optimal recovery strategy for a damaged network varies with the extent of the damage. From the results of the simulation of the attack-recovery process, we conclude that while defining and analyzing complex network models, we should adjust our network topology, weight assignment, and performance indicators in accordance with the focal characteristics of complex systems so that we can use the network model to build robust complex systems and efficient logistics and maintenance strategies.

Quantitative method for resilience assessment framework of airport network during COVID-19

“…Nodes were removed based on different weight indicators. We removed nodes in the order of decreasing node degree, betweenness centrality, and node strength [31].…”

Resilience based importance measure analysis for SoS

“…Network efficiency and largest connected component (LCC) are two common performance measurements in the study of complex networks. To some extent, these traditional measurements reflect the performance of the SoS, while they are limited to some certain aspects of the performance [38]. For example, network efficiency focuses more on the ability of the network to transmit information, while LCC neglects the functionality of other connected groups of nodes.…”

Topological Approach to Measure Network Recoverability

“…In a board sense, network robustness is also related to the ability of a network to return to a desired performance level after suffering malicious attacks and random failures [8]. We define such network capability as network recoverability 1 in this paper.…”

“…Majdandzic et al [10] model cascading failures and spontaneous recovery as a stochastic contiguous spreading process and exhibit a phase switching phenomenon. The recovery strategies based on the centrality metrics of network elements (e.g., nodes or links) are investigated in [8] [11], which show that a centrality metric-based strategy may not exist to improve all the network performance aspects simultaneously. A progressive recovery approach [12], that consists in choosing the right sequence of links to be restored after a disaster in communication networks, proposes to maximize the weighted sum of the total flow over the entire process of recovery [13], as well as to minimize the total cost of repair under link capacity constraints [14].…”

“…Although the above papers [8]- [14] have contributed to this field, a general framework or methodology for quantifying the recovery capability of a real communication network is still lacking. In this paper, we propose a topological approach and two recoverability indicators to measure the network recoverability in two different scenarios, link-based Scenario A and energy-based Scenario B.…”