Project

Evolution and learning in strategic games have been a very active field of investigation in the last decade. The reason is the convergence of the newly founded science of complex networks on one side, and the use of sophisticated models for interacting strategic agents on the other. In this project we continue our research in the field and extend it further to areas that have not been investigated enough or at all up to now. In particular, we propose to study iterated games, i.e. games in which a group of agents interact repeatedly,  on static and dynamically changing complex networks and to investigate more sophisticated forms of agent behavior in populations based on machine learning ideas such as reinforcement learning. Moreover, following a recent strand of our own research, we shall continue experimental work on simple games in the laboratory, extending it to dynamical cases in which the interaction between agents may evolve in time. This aspect of the project is important because it would allow us to compare real players behavior and various learning models with the goal of improving the latter on one hand, and of suggesting interesting experimental settings by using numerical simulations of the models on the other hand.