Selection Schemes Analysis in Genetic Algorithms for the Maximum Influence Problem

Information spread in social network is a current prime target for a number of sectors, namely politics, marketing, research, education, finance, etc. Information diffusion through the network has been modeled in different manners, all of them using their own dynamics. The main goal is to maximize the influence with the minimum number of starting users. This problem is known as the influence maximization problem, which is known to be NP-hard. This is why several proposals based on heuristics and meta-heuristics have appeared in order to tackle the problem. Interesting results have been published, however, many studies have concentrated exclusively on the results and the analysis of the algorithms components has been left aside. We believe it is also important to know what features of the algorithms are meaningful in order for the algorithms to perform well. This is why we analyze a couple of selection schemes in a genetic algorithm. Our results revealed that one of the selection schemes perform better for a certain class of networks.