Properties of Networks of Interacting Stochastic Agents
This event is part of the Departmental Seminars.
Eugene Stanley, Irena Vodenska, Shlomo Havlin, Pankaj Mehta, and Kevin Black.
Networks generally represent a type of interaction between entities. Sometimes these interactions encode little about the similarities of agents and it is more the fame that determines links, as is the case in Twitter. But in the formation of many other networks similarity is crucial. The similarities may include overlap of interests, field of study, or geographical location. Each of these cases may be thought of as dimensions of a parameter space. Similarity then becomes locality and local interactions inside this parameter space. We devise a general model for formation of such locality-based networks using field theory and show that the locality puts clear constraints on the structure of the network. The field theory machinery allows us find analytic relations for network characteristics. We show that some very generic cases of networks produced from our model can reproduce features of many similarity-based real-world networks and contrast them with non-local networks.