| Abstract |
While there exists considerable work that looks into optimally influencing individuals within a given network to maximise the overall spread of influence, existing work in this area typically neglects a few key features present in realistic application domains: (i) individuals may change their opinions dynamically over time, (ii) varying amounts of resources (or incentives) may be expended on individuals to influence their opinions. In this paper, we address these shortcomings. Specifically, using the principles of game theory and voter dynamics, we propose a computational model that solves this optimisation problem of allocating continuous resources within a given budget to maximise influence spread in the presence of an adversary (whose strategy may be known or unknown) in a canonical star topology. |