An Adversarial Dynamic Game to Controlling Information Diffusion under Typical Strategies on Online Social Networks

Abstract

The diffusion of negative information, such as rumours, misinformation and computer viruses on Online Social Networks (OSNs), may lead to serious losses and consequences. And there are always some rational malicious spreaders, who strategically spread negative information. Therefore, how to control the information diffusion of the malicious spreader is a great challenge. In recent years, some studies have analyzed the controlling problem which belongs to the issue of influence blocking maximization (IBM) from the perspective of the large-scale strategy set on the game theory. However, the aforementioned methods cannot timely solve the controlling diffusion problem on high-speed OSNs. In this study, we achieve the purpose of effectively controlling diffusion on OSNs by blocking information under typical strategies. Based on the existing two-player Stackelberg zero-sum game model and evaluation methods of node’s importance on the network, we analyze the typical strategic dynamic game in which the blocker moves first and the spreader moves later on scale-free networks with different power exponent. Experimental results show that the absolute dominance strategy of the blocker is Leader Rank with 90.16% probability. And using Leader Rank can be relatively effective against malicious spreaders with 98.33% probability. When the power exponent of the network is smaller, it is more conducive to blocking information dissemination with fewer seed nodes.