Stability of Gradient Learning Dynamics in Continuous Games: Vector Action Spaces

by   Benjamin J. Chasnov, et al.

Towards characterizing the optimization landscape of games, this paper analyzes the stability of gradient-based dynamics near fixed points of two-player continuous games. We introduce the quadratic numerical range as a method to characterize the spectrum of game dynamics and prove the robustness of equilibria to variations in learning rates. By decomposing the game Jacobian into symmetric and skew-symmetric components, we assess the contribution of a vector field's potential and rotational components to the stability of differential Nash equilibria. Our results show that in zero-sum games, all Nash are stable and robust; in potential games, all stable points are Nash. For general-sum games, we provide a sufficient condition for instability. We conclude with a numerical example in which learning with timescale separation results in faster convergence.



page 1


Stability of Gradient Learning Dynamics in Continuous Games: Scalar Action Spaces

Learning processes in games explain how players grapple with one another...

Convergence of Learning Dynamics in Stackelberg Games

This paper investigates the convergence of learning dynamics in Stackelb...

Local Nash Equilibria are Isolated, Strict Local Nash Equilibria in `Almost All' Zero-Sum Continuous Games

We prove that differential Nash equilibria are generic amongst local Nas...

Nash Equilibria of The Multiplayer Colonel Blotto Game on Arbitrary Measure Spaces

The Colonel Blotto Problem proposed by Borel in 1921 has served as a wid...

Robustness against Agent Failure in Hedonic Games

We study how stability can be maintained even after any set of at most k...

Disturbance Decoupling for Gradient-based Multi-Agent Learning with Quadratic Costs

Motivated by applications of multi-agent learning in noisy environments,...

Gains in evolutionary dynamics: a unifying approach to stability for contractive games and ESS

In this paper, we investigate gains from strategy revisions in determini...
This week in AI

Get the week's most popular data science and artificial intelligence research sent straight to your inbox every Saturday.