Influncing Opinion evolution over finite time horizons

Abstract

We consider an extension of the voter model and study the evolution of a fixed as well as a growing population over a finite time horizon, in presence of an external influencing agency. Opinion models with external influence are increasingly relevant to the current form of social interactions, especially on online social media platforms. Short term opinion polls and focused advertising are fairly common phenomena on these social networking sites. Building on the existing work for a constant population model, we obtain optimal strategies for the influencing agency (with limited budget) in terms of when to influence. In other words, given that the external agency can influence only a fraction of total time-slots, depending on the behaviour of the individuals in the population, we determine which time- slots should be influenced in order to maximise the expected fraction of favourable opinion at the end of the time horizon.