The role of convexity in saddle-point dynamics: Lyapunov function and robustness
نویسندگان
چکیده
This paper studies the projected saddle-point dynamics associated to a convex-concave function, which we term saddle function. The dynamics consists of gradient descent of the saddle function in variables corresponding to convexity and (projected) gradient ascent in variables corresponding to concavity. We examine the role that the local and/or global nature of the convexity-concavity properties of the saddle function plays in guaranteeing convergence and robustness of the dynamics. Under the assumption that the saddle function is twice continuously differentiable, we provide a novel characterization of the omegalimit set of the trajectories of this dynamics in terms of the diagonal blocks of the Hessian. Using this characterization, we establish global asymptotic convergence of the dynamics under local strong convexity-concavity of the saddle function. When strong convexity-concavity holds globally, we establish three results. First, we identify a Lyapunov function (that decreases strictly along the trajectory) for the projected saddle-point dynamics when the saddle function corresponds to the Lagrangian of a general constrained convex optimization problem. Second, for the particular case when the saddle function is the Lagrangian of an equality-constrained optimization problem, we show inputto-state stability of the saddle-point dynamics by providing an ISS Lyapunov function. Third, we use the latter result to design an opportunistic state-triggered implementation of the dynamics. Various examples illustrate our results.
منابع مشابه
Estimation of the Domain of Attraction of Free Tumor Equilibrium Point for Perturbed Tumor Immunotherapy Model
In this paper, we are going to estimate the domain of attraction of tumor-free equilibrium points in a perturbed cancer tumor model describing the tumor-immune system competition dynamics. The proposed method is based on an optimization problem solution for a chosen Lyapunov function that can be casted in terms of Linear Matrix Inequalities constraint and Taylor expansion of nonlinear terms. We...
متن کاملPeaking Attenuation in High-Gain Observers Using Adaptive Saturation: Application to a Ball and Wheel System
Despite providing robustness, high-gain observers impose a peaking phenomenon, which may cause instability, on the system states. In this paper, an adaptive saturation is proposed to attenuate the undesirable mentioned phenomenon in high-gain observers. A real-valued and differentiable sigmoid function is considered as the saturating element whose parameters (height and slope) are adaptively tu...
متن کاملSaddle-Point Dynamics: Conditions for Asymptotic Stability of Saddle Points
This paper considers continuously differentiable functions of two vector variables that have (possibly a continuum of) min-max saddle points. We study the asymptotic convergence properties of the associated saddle-point dynamics (gradient-descent in the first variable and gradientascent in the second one). We identify a suite of complementary conditions under which the set of saddle points is a...
متن کاملStability and Bifurcation of an SIS Epidemic Model with Saturated Incidence Rate and Treatment Function
In this paper an SIS epidemic model with saturated incidence rate and treatment func- tion is proposed and studied. The existence of all feasible equilibrium points is discussed. The local stability conditions of the disease free equilibrium point and endemic equilibrium point are established with the help of basic reproduction number.However the global stabili- ty conditions of these eq...
متن کاملLyapunov-Based Robust Power Controllers for a Doubly Fed Induction Generator
In this work, a robust nonlinear control technique of a doubly fed induction generator (DFIG) intended for wind energy systems has been proposed. The principal idea in this article is to decouple the active and reactive power of the DFIG with high robustness using the backstepping strategy. The principle of this control method is based on the Lyapunov function, in order to guarantee the global ...
متن کامل