MECIM: Multi-entity evolutionary competitive influence maximization in social networks

Competitive Influence Maximization (CIM) is a crucial problem in social networks to identify a small set of nodes in competitive environments that possess the potential to maximize the spread of influence. Despite some research efforts to solve the CIM problem, they fail to consider the uncertainty of data during transmission, limiting the spread of influence in practical scenarios. In this paper, we formulate the Multi-entity Evolutionary Competitive Influence Maximization (MECIM) problem, which aims to maximize the spread of influence in competitive social networks by considering the uncertainty of data during transmission. To tackle this problem, we propose the Multi-entity Evolutionary Competitive Independent Cascade (MECIC) model. This model introduces multiple node states to handle the uncertainty of neighbor node states and integrates the dynamic and competitive characteristics of networks to enhance the accuracy of identifying seed nodes. Moreover, we propose three algorithms to maximize the influence spread. Theoretical results are provided to demonstrate the efficiency of our design. Finally, extensive experiments conducted on real-world datasets indicate that the proposed algorithms efficiently maximize the influence spread over competitive social networks.