An improved competitive particle swarm optimization algorithm based on de-heterogeneous information

Abstract

The problem of influence dissemination under competitive conditions can truly reflect the form of information in social networks, make reasonable decisions, and analyze the impact gain of different types of messages on users, so as to maximize the benefits of users. How to accurately simulate the spread of competitive influence is still a challenging problem. In such conditions to carry out the information dissemination, viral marketing, disease prevention and other works has a more reasonable significance. However, it is NP-hard to find the most influential node set in a competitive environment. Although an approximate optimal solution can be obtained by using a greedy algorithm to simulate influence propagation under competitive conditions, the time efficiency of the process is very low. Therefore, this paper designs a bionic optimized solution to explore the above problems. Firstly, We set optimization measures to balance the convergence and diversity of the optimization algorithm in the target space, and based on this, we propose a local impact ability evaluation metric based on the de-heterogeneous information within the 2-hop neighborhood of the node set to measure the propagation ability of the node set under competitive conditions. Secondly, the proposed evaluation index is applied to the improved particle swarm optimization algorithm to realize the screening of global optimal particles. Finally, we build a game model to simulate the interest-driven information ownership and dissemination of nodes. The effectiveness and efficiency of our method are verified by comparing the propagation range of competitive influence of different methods and the time complexity of each algorithm on real data sets.

Keywords

• Maximizing competitive influence
• De-heterogeneous information
• Particle swarm algorithm
• Game model