Network games, equilibrium learning and intervention
keywords COMPLEX SYSTEMS, GAME THEORY, INCENTIVES, SOCIAL NETWORKS
Reference persons LUCA DALL'ASTA
Research Groups AA - Statistical Physics and Interdisciplinary Applications
Description Network games are stylized strategic problems in which individual utilities are affected by local externalities on a given interaction network. Depending on the level of knowledge agents have on the local/global structure of the network, different solution concepts have been proposed to describe collective rational behavior. For instance, self-confirming equilibria are generalizations of Nash equilibria in which a form of consistency between subjective utility functions and individual beliefs based on repeated observations is assumed. Such equilibria play an important role in the context of game-theoretic learning processes. The thesis work is focused on the development of statistical physics methods to study the properties of self-confirming equilibria on networks, both from a static and dynamic point of view, and tackle the problem of optimal incentive allocation in network games.
Required skills Knowledge of fundamental concepts and methods in game theory, dynamical systems, and graph theory. Good programming skills (e.g. Python/Julia/C++).
Deadline 02/02/2024 PROPONI LA TUA CANDIDATURA