Stability and instability in saddle point dynamics -- Part I

TL;DR

This paper analyzes the long-term behavior of gradient and subgradient dynamics in saddle point problems, providing explicit linear ODE characterizations and highlighting potential issues with convergence and stability.

Contribution

It offers an exact characterization of the asymptotic behavior of unconstrained gradient dynamics for concave-convex functions, including conditions for convergence and instability.

Findings

Asymptotic behavior is described by solutions to explicit linear ODEs.

Non-convergence can lead to unbounded variance in system behavior.

Results inform convergence criteria and stability analysis.

Abstract

We consider the problem of convergence to a saddle point of a concave-convex function via gradient dynamics. Since first introduced by Arrow, Hurwicz and Uzawa in [1] such dynamics have been extensively used in diverse areas, there are, however, features that render their analysis non trivial. These include the lack of convergence guarantees when the function considered is not strictly concave-convex and also the non-smoothness of subgradient dynamics. Our aim in this two part paper is to provide an explicit characterization to the asymptotic behaviour of general gradient and subgradient dynamics applied to a general concave-convex function. We show that despite the nonlinearity and non-smoothness of these dynamics their $\omega$-limit set is comprised of trajectories that solve only explicit linear ODEs that are characterized within the paper. More precisely, in Part I an exact characterization is provided to the asymptotic behaviour of unconstrained gradient dynamics. We also show that when convergence to a saddle point is not guaranteed then the system behaviour can be problematic, with arbitrarily small noise leading to an unbounded variance. In Part II we consider a general class of subgradient dynamics that restrict trajectories in an arbitrary convex domain, and show that when an equilibrium point exists their limiting trajectories are solutions of subgradient dynamics on only affine subspaces. The latter is a smooth class of dynamics with an asymptotic behaviour exactly characterized in Part I, as solutions to explicit linear ODEs. These results are used to formulate corresponding convergence criteria and are demonstrated with several examples and applications presented in Part II.