EFFECT OF ZEALOTS ON THE OPINION DYNAMICS OF RATIONAL AGENTS WITH BOUNDED CONFIDENCE

This paper incorporates a micro-level decision-making paradigm along with a social interaction model (bounded confidence) in the presence of influences (zealots). Every agent in the society represents a node in a Barabasi-Albert network and is given a decision-making ability (to choose from a fixed set of states). The decision making is based on maximization of estimated accumulated rewards gained as a result of an individual's own sequence of choices in the presence of different probabilities of external events. The effects of interactions, and events on the final distribution of decision states are studied with and without the presence of influences. Bounded confidence model parameters (the distance parameter and the convergence parameter) are used to study the final distribution of states, and the time the society needs to reach its equilibrium (convergence time). Finally, effects of network topology on the final distribution of states and convergence time are presented.