Exploring Social Phenomena with Complex Systems Tools.

Summary The main objective of the thesis is to explore problems specific to the social sciences and to study them using tools from the field of statistical physics and complex systems. The work of the thesis is thus declined on three main topics whose main problem is the question of the aggregation of individual entities into a collective structure. The first topic is centered on a paradigmatic example of the emergence of an unexpected macroscopic collective behavior from simple individual rules: the Schelling segregation model. In particular, we have proposed a novel analytical resolution of this model and we have studied analytically and via simulations the impact of different forms of cooperation between individual agents on the global collective behavior. This topic has been developed both from an economic and a physical point of view. The second topic concerns the exploration of bibliometric databases. We have produced 'science maps' representing the field of complex systems (its internal structure being deciphered via an analysis of the references used by ~1,400,000 articles) or the state of research within an institution such as the ENS de Lyon. Finally, the third theme deals with the development of models based on 'hard' science tools but sociologically founded. We present the process of elaboration of a model built with a team of sociologists. Finally, we develop a model of opinion that specifically answers a question: the existence of structures that last from entities that do not last.