Improving robustness of complex networks by a new capacity allocation strategy

Abstract

The robustness of infrastructure networks has attracted great attention in recent years. Scholars have studied the robustness of complex networks against cascading failures from different aspects. In this paper, a new capacity allocation strategy is proposed to reduce cascading failures and improve network robustness without changing the network structure. Compared with the typical strategy proposed in Motter–Lai (ML) model, the new strategy can reduce the scale of cascading failure. The new strategy applied in scale-free network is more efficient. In addition, to reasonably evaluate the two strategies, we introduce contribution rate of unit capacity to network robustness as evaluation index. Results show that our new strategy works well, and it is more advantageous in the rational utilization of capacity in scale-free networks. Furthermore, we were surprised to find that the efficient utilization of capacity costs declined as costs rose above a certain threshold, which indicates that it is not wise to restrain cascading failures by increasing capacity costs indefinitely.

Robustness of community networks against cascading failures with heterogeneous redistribution strategies

“…[6][7][8][9] For example, the heterogeneity of nodes and communities in networks has an important impact on the whole cascading failure process, and node loadredistribution strategies based on node's influence are continuously proposed. [3,4,10] Therefore, the impact of network structure on cascading failure has been one of the hot topics in research of complex network dynamics.…”

Effects of heterogeneous adoption thresholds on contact-limited social contagions

“…[1][2][3][4][5][6][7][8][9] A real complex system can be abstracted as a network, and then complex properties of the system, such as structure, robustness, function, as well as dynamical processes that occur on networks, can be investigated. [10][11][12][13][14][15][16][17][18][19][20][21][22][23] According to different characteristics of spreading objects, the spreading dynamics on complex networks can be generally divided into two categories: one is the dynamics of biological contagions with independent infection probabilities of any two contacts, such as the epidemic spreading and the spreading of computer viruses, [24][25][26][27][28][29][30] the other is the dynamics of social contagions with social reinforcement effect, including rumor diffusion, information spreading, behavior spreading, innovation adoption, and so on. [31][32][33][34][35][36][37][38] It is generally believed that social contagion is the expansion and application of biological contagion, and it is the research basis of social-biological coupling contagion.…”

Detection of influential nodes with multi-scale information*

“…[7][8][9] As one of the most significant aspects of the study of complex networks, identifying the nodes with the maximum influence plays an essential role in various practical applications, such as controlling the spread of an epidemic, [10] ranking web pages, [11] forecasting economic trends, [12] detecting of centrality nodes in community partition [13][14][15] and improving the robustness of networks. [16][17][18] Therefore, research on identifying influential nodes in a given network becomes an important topic.…”

“…There are numerous methods for evaluating the importance of nodes in a network, such as degree centrality, [19] the neighbor nodes degree algorithm (NND), [20] closeness centrality, [21] betweenness centrality, [22,23] Katz centrality, [24] eigenvector centrality, [25] PageRank, [26] the k-shell decomposition method [27] and many others. [28][29][30][31] Among these, degree centrality [18] is of the simplest measures that only considers local topology information and counts the number of neighbors of the node. The NND algorithm [20] considers both the node degree and the node neighbor's degree.…”