Modeling Leadership Hierarchy in Multilevel Animal Societies

Abstract

A typical feature of many natural and social networks is the presence of communities giving rise to multiple levels of organization. We investigate the decision-making process of a group combining self organization and social dynamics, and reproduce the simultaneous emergence of a hierarchical and modular leadership network. All individuals in the model try, with varying degrees of ability, to find a direction of movement, with the result that leader-follower relationships evolve between them, since they tend to follow the more successful ones. The harem-forming ambitions of male individuals inspired by an observed Przewalski horse herd (Hortobagy, Hungary) leads to modular structure. In this ´ approach we find that the harem-leader to harem-member ratio observed in horses corresponds to an optimal network regarding common success, and that modularly structured hierarchy is more benefical than a non-modular one, in the sense that common success is higher, and the underlying network is more hierarchical. We also find that the experimental and model harem size distributions are close to a lognormal.