Maximizing Influence in Network Dynamics with Social Considerations

This thesis explores the problem of influence maximization in social networks, with a focus on algorithms that take into account the existence of vulnerable users. We begin by modeling social networks and examining foundational properties, such as submodu- larity of the spread function. Various algorithms, including Naïve Greedy, subsampling and sandwich approximation, are reviewed for their efficiency and applicability. The core of our study examines some classic as well as innovative strategies like Differ- ence Maximization, Ratio Maximization and Additive Smoothing Ratio Maximization, alongside a PageRank-inspired method to balance influence spread and protection of vulnerable users. Our findings aim to provide solutions that optimize influence while adopting a more socially responsible approach.