Distributed Nash Equilibrium Seeking of A Class of Aggregative Games

In this paper, we investigate a distributed Nash equilibrium seeking problem for a class of aggregative games that the strategic interaction is characterized by a sum of nonlinear mapping of heterogeneous local decisions. We consider non-quadratic local cost functions and constrained strategy sets. We propose a novel continuous-time distributed algorithm for equilibrium seeking based on dynamic average consensus protocol and projected gradient flows. By virtue of variational inequalities and Lyapunov functions, we show the convergence of our distributed algorithm for the aggregative game.