| KEYWORD |
TARGETING INTERVENTIONS IN NETWORK GAMES (OF STRATEGIC COMPLEMENTS)
keywords GAME THEORY, INTERVENTION, NETWORK DYNAMICS
Reference persons GIACOMO COMO, FABIO FAGNANI
Research Groups Analisi e controllo di sistemi dinamici
Description How much energy does it take to modify the behavior of a complex system? Which nodes of a network, under the effect of an exogenous intervention, maximally affect the system as a whole, possibly triggering cascade effects? These questions are ubiquitous and of paramount importance in a number of scientific disciplines focused on the study of networks. In particular, examples come from social sciences and economics in the study of the diffusion of new habits, technologies and new ideas in a population.
One popular model to describe the interactions among rational agents is that of game theory. In such models players are endowed with a utility function that depends on the action they choose to play and on the actions played by the other players. We talk about network games when the strategic interactions among players are described by a graph so that each player only interacts with its neighboring players. A crucial concept in this field is that of Nash equilibrium that is a configuration of actions such that no player can improve its utility by unilaterally modifying its own action. Under certain assumptions natural learning dynamical rules lead the system to converge to a Nash equilibrium. An intervention strategy consists in modifying the utility of a family of targeted players in order to induce desirable behaviors in the system. This might consist in shaping the family of the Nash equilibria or rather in pushing certain learning dynamics to converge towards a specific Nash equilibrium that from the point of view of an external planner represents the optimal social choice.
Further details can be found in the attached file.
See also tesi1.pdf
Deadline 31/12/2021
PROPONI LA TUA CANDIDATURA