Load Distribution-Based Research on a Cascading Failure Model of Network Public Opinion Dissemination

Abstract

Research of network cascading failures is a heavily researched topic in network science research. The expression of a network cascading failure mechanism by a dynamic equation has become a basic mode for research of cascading failures. With the promotion of complexity and intelligent of nodes, it is difficult to reflect the evolution of the actual large network state objectively that only considering the coupling relationship among nodes from one network. To describe the dissemination process of network public opinion more accurately, this paper proposes a failure propagation model of network public opinion dissemination based on load spatial distribution, in which evolutional behaviors of nodes were expanded to overall evolutional behaviors of a network based on network coupling relations, and difficulty of analysis and solution was reduced through periodic iteration and feature value decomposition of network adjacent matrixes. Experimental results verify the feasibility and validity of this method. Failure analysis of the computer network also indicates that node behaviors could facilitate weakening of effects generated from attacks and failures. Meanwhile, weakening of nodes could effectively reduce the network sensitivity to deliberate attacks.

Keywords

• Public opinion
• cascading failure
• differential dynamic system
• load distribution
• propagation model