Logo image
Equilibria and bifurcation theory for mean-field games
Dissertation   Open access

Equilibria and bifurcation theory for mean-field games

Luke Candler Brown
Doctor of Philosophy (Ph.D.), Drexel University
Jun 2023
DOI:
https://doi.org/10.17918/00001649
pdf
Brown_Luke_20231.76 MBDownloadView

Abstract

Fokker-Planck equation Hamilton-Jacobi equations Mean field theory Hamilton-Jacobi-Bellman Mean-field games PDE Partial differetial equations Economics
To represent the interaction of N rational competitors traditionally, a coupled system of N differential equations must be solved simultaneously, yielding the equilibrium strategy for each player. This approach becomes impractical as N grows, prompting the adoption of a mean-field approach, in which we assume N is large enough that the dynamics of the competition may be suitably represented distributionally in the continuum case. The trajectory of each player through the state space is then driven by a dynamic control, together with an adapted Brownian motion, reducing the computation from N differential equations to two Partial Differetial Equations (PDE's), a Hamilton-Jacobi-Bellman equation governing the evolution in time of a utility function, and a Fokker-Planck (Forward Kolmogorov) equation governing the evolutio n in time of the distribution of players, coupled by means of the nonlinear Hamiltonian. In this thesis, we use the Implicit Function Theorem and Bifurcation Theory to obtain nontrivial equilibria of various mean-field games, and we go on to demonstrate the use of Schauder's fixed-point theorem to prove the existence of low-regularity time-bound solutions of a congestion-type mean-field game.

Metrics

84 File views/ downloads
120 Record Views

Details

Logo image