Evolutionary-based Framework for Optimizing the Spread of Information on Twitter

Currently social networks are a medium designed to share and spread information and opinions among users. To effectively spread information characteristics of information sources and their positions in the network have to be properly adjusted. Sources play a crucial role as they initiate and support the spreading process. Adjusting sources has a certain cost and thus their non-optimal configuration may lead to wasting resources spent to sources adjusting or even poor performance. In this work, the problem of sources settings and positions optimization is discussed and the framework for solving the problem is proposed. The developed framework incorporates source and spreading models which take into account individual characteristics of sources as characteristics of their positions. The solution can handle all steps which may be required for optimization: network monitoring and data collecting; parameters and settings identification; source layout optimization. The efficiency of solutions generated by the framework with the genetic algorithm adapted for source layout optimization is demonstrated by comparison with greedy heuristic on the dataset collected from Twitter.