Influence Maximization in Network by Genetic Algorithm on Linear Threshold Model

摘要

The problem of maximum influence on the network consists in the search for a subset of k vertices called seeds which when activated are able to influence as much elements as possible, considering a model to simulate the propagation of influence in a network. This paper proposes a Genetic Algorithm to optimize the selection of seeds for the Linear Threshold Model (LTM), a widely adopted simulation model for influence propagation, by investigating different strategies for initial population configurations based on high centrality nodes. The results obtained by the application of the proposed methodology to the Linear Threshold Model considering real world networks show significant improvements on the convergence of the algorithm.