基于库仑力模型的动态社会网络积极影响力最大化算法

Dynamic social network active influence maximization algorithm based on Coulomb force model

影响力最大化问题已经成为社会网络中重要的研究内容,其影响力传播模型和求解算法是关键的核心问题。为了提高预测传播结果的准确度,引入传播过程中激活节点数量动态变化与节点间信任关系对IC模型进行改进,结合社会影响力与库仑力之间的相似性,提出一种基于信任关系的动态社会库仑力(dynamic social Coulomb forces based on trust relationship,DSC-TR)模型,构建一种优化的随机贪心(random greedy,RG-DPIM)算法求解影响最大化问题。仿真实验结果表明,DSC-TR模型的预测准确度明显优于SC-B、IC模型;RG-DPIM算法性能优于G-DPIM、IPA、TDIA算法。

The problem of maximizing influence has become an important research content in social networks,and its influence propagation model and solving algorithm are the key core issues.In order to improve the accuracy of predicting the propagation results,the dynamic change of the number of activated nodes and the trust relationship between the nodes during the propagation process were introduced to improve the IC model.Combining the similarity between social influence and Coulomb force,a dynamic based on trust relationship was proposed,a dynamic social coulomb forces based on trust relationships(DSC-TR)model was proposed,and an optimized random greedy(RG-DPIM)algorithm was constructed to solve the problem of maximum impact.Simulation results show that the prediction accuracy of the DSC-TR model is obviously better than that of SC-B and IC models.The performance of RG-DPIM algorithm is obviously better than that of G-DPIM,IPA and TDIA algorithms.