arXiv subjects

Jiangxing Zhu

Publications and source records attributed to Jiangxing Zhu.

4 recordsLinked to original sources

arXiv2024

Generalized Metric Subregularity with Applications to High-Order Regularized Newton Methods

This paper pursues a twofold goal. First, we introduce and study in detail a new notion of variational analysis called generalized metric subregularity, which is a far-going extension of the conventional metric subregularity conditions. Our primary focus is on examining this concept concerning first-order and second-order stationary points. We develop an extended convergence framework that enables us to derive superlinear and quadratic convergence under the generalized metric subregularity condition, broadening the widely used KL convergence analysis framework. We present verifiable sufficient conditions to ensure the proposed generalized metric subregularity condition and provide examples demonstrating that the derived convergence rates are sharp. Second, we design a new high-order regularized Newton method with momentum steps, and apply the generalized metric subregularity to establish its superlinear convergence. Quadratic convergence is obtained under additional assumptions. Specifically, when applying the proposed method to solve the (nonconvex) over-parameterized compressed sensing model, we achieve global convergence with a quadratic local convergence rate towards a global minimizer under a strict complementarity condition.

math.OC↗

arXiv2021

Isolated calmness and sharp minima via Hölder graphical derivatives

The paper utilizes Hölder graphical derivatives for characterizing Hölder strong subregularity, isolated calmness and sharp minimum. As applications, we characterize Hölder isolated calmness in linear semi-infinite optimization and Hölder sharp minimizers of some penalty functions for constrained optimization.

math.OC↗

arXiv2019

Hölder Error Bounds and Hölder Calmness with Applications to Convex Semi-Infinite Optimization

Using techniques of variational analysis, necessary and sufficient subdifferential conditions for Hölder error bounds are investigated and some new estimates for the corresponding modulus are obtained. As an application, we consider the setting of convex semi-infinite optimization and give a characterization of the Hölder calmness of the argmin mapping in terms of the level set mapping (with respect to the objective function) and a special supremum function. We also estimate the Hölder calmness modulus of the argmin mapping in the framework of linear programming.

math.OC↗

arXiv2016

Stable Well-posedness and Tilt stability with respect to admissible functions

Note that the well-posedness of a proper lower semicontinuous function $f$ can be equivalently described using an admissible function. In the case when the objective function $f$ undergos the tilt perturbations in the sense of Poliquin and Rockafellar, adopting admissible functions $φ$ and $ψ$, this paper introduces and studies the stable well-posedness of $f$ with respect to $φ$ (in breif, $φ$-SLWP) and tilt-stable local minimum of $f$ with respect to $ψ$ (in brief, $ψ$-TSLM). In the special case when $φ(t)=t^2$ and $ψ(t)=t$, the corresponding $φ$-SLWP and $ψ$-TSLM reduce to the stable second local minimizer and tilt stable local minimum respectively, which have been extensively studied in recent years. We discover an interesting relationship between two admissible functions $φ$ and $ψ$: $ψ(t)=(φ')^{-1}(t)$, which implies that a proper lower semicontinous function $f$ on a Banach space has $φ$-SLWP if and only if $f$ has $ψ$-TSLM. Using the techniques of variational analysis and conjugate analysis, we also prove that the strong metric $φ'$-regularity of $\partial f$ is a sufficient condition for $f$ to have $φ$-SLWP and that the strong metric $φ'$-regularity of $\partial\overline{\rm co}(f+δ_{B[\bar x,r]})$ for some $r>0$ is a necessary condition for $f$ to have $φ$-SLWP. In the special case when $φ(t)=t^2$, our results cover some existing main results on the tilt stability.

math.OC↗