Convergence Analysis of Gradient Algorithms on Riemannian Manifolds without Curvature Constraints and Application to Riemannian Mass

DOI10.1137/19M1289285MaRDI QIDQ5148408zbMATH OpenOpenAlexWikidataFDO

Authors Xiangmei Wang, C. Li, J. H. Wang, Jen-Chih Yao

Publication date 4 February 2021

Published in SIAM Journal on Optimization (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1910.02280

zbMATH Keywords

global convergencegradient algorithmlinear convergencelocal convergenceRiemannian center of massRiemannian manifoldsectional curvature

Mathematics Subject Classification ID

Global Riemannian geometry, including pinching (53C20) Geodesics in global differential geometry (53C22) Mathematical programming (90C99)

Abstract: We study the convergence issue for the gradient algorithm (employing general step sizes) for optimization problems on general Riemannian manifolds (without curvature constraints). Under the assumption of the local convexity/quasi-convexity (resp. weak sharp minima), local/global convergence (resp. linear convergence) results are established. As an application, the linear convergence properties of the gradient algorithm employing the constant step sizes and the Armijo step sizes for finding the Riemannian $L^p$ ($pin[1,+infty)$) centers of mass are explored, respectively, which in particular extend and/or improve the corresponding results in cite{Afsari2013}.

Recommendations

• Riemannian conjugate gradient methods: general framework and specific algorithms with convergence analyses
• scientific article; zbMATH DE number 938816
• On the convergence of gradient descent for finding the Riemannian center of mass
• Gradient method for optimization on Riemannian manifolds with lower bounded curvature
• Convergence of inexact steepest descent algorithm for multiobjective optimizations on Riemannian manifolds without curvature constraints
• scientific article; zbMATH DE number 4031421
• The study of the subgradient algorithm and its convergence on Riemannian manifolds
• A Riemannian conjugate gradient method for optimization on the Stiefel manifold
• Linear convergence of subgradient algorithm for convex feasibility on Riemannian manifolds
• Globally convergent optimization algorithms on Riemannian manifolds: Uniform framework for unconstrained and constrained optimization

Cites work

• A.D. Alexandrov spaces with curvature bounded below
• Characterizations of strict local minima and necessary conditions for weak sharp minima
• Computing Riemannian center of mass on Hadamard manifolds
• Convergence analysis of inexact proximal point algorithms on Hadamard manifolds
• Existence of solutions for variational inequalities on Riemannian manifolds
• Geometry and spectra of compact Riemann surfaces
• Globally convergent optimization algorithms on Riemannian manifolds: Uniform framework for unconstrained and constrained optimization
• Gradient method for optimization on Riemannian manifolds with lower bounded curvature
• scientific article; zbMATH DE number 5704332 (Why is no real title available?)
• scientific article; zbMATH DE number 1376185 (Why is no real title available?)
• scientific article; zbMATH DE number 1282147 (Why is no real title available?)
• scientific article; zbMATH DE number 681023 (Why is no real title available?)
• scientific article; zbMATH DE number 909255 (Why is no real title available?)
• scientific article; zbMATH DE number 3214233 (Why is no real title available?)
• scientific article; zbMATH DE number 5223994 (Why is no real title available?)
• Linear convergence of subgradient algorithm for convex feasibility on Riemannian manifolds
• Locally Geodesically Quasiconvex Functions on Complete Riemannian Manifolds
• Minimizing a differentiable function over a differential manifold
• Monotone and accretive vector fields on Riemannian manifolds
• Monotone vector fields and the proximal point algorithm on Hadamard manifolds
• Newton's method on Riemannian manifolds and a geometric model for the human spine
• Newton's method on Riemannian manifolds: Smale's point estimate theory under the γ-condition
• Newton's method, zeroes of vector fields, and the Riemannian center of mass
• On the convergence of gradient descent for finding the Riemannian center of mass
• Optimal stability and eigenvalue multiplicity
• Quotient geometry with simple geodesics for the manifold of fixed-rank positive-semidefinite matrices
• Riemannian \(L^{p}\) center of mass: existence, uniqueness, and convexity
• Riemannian Geometry
• Riemannian median and its estimation
• Second-order Sufficiency and Quadratic Growth for Nonisolated Minima
• Steepest descent method with a generalized Armijo search for quasiconvex functions on Riemannian manifolds
• Stochastic algorithms for computing means of probability measures
• Subgradient algorithms on Riemannian manifolds of lower bounded curvatures
• Subgradient projection algorithms for convex feasibility on Riemannian manifolds with lower bounded curvatures
• The Gradient Projection Method Along Geodesics
• Trust-region methods on Riemannian manifolds
• Unconstrained steepest descent method for multicriteria optimization on Riemannian manifolds
• Variational inequalities for set-valued vector fields on Riemannian manifolds: convexity of the solution set and the proximal point algorithm
• Weak Sharp Minima in Mathematical Programming
• Weak sharp minima on Riemannian manifolds
• Weak sharp minima revisited. I: Basic theory
• Weak Sharp Minima: Characterizations and Sufficient Conditions

Cited in

(20)

• Levitin-Polyak well-posedness by perturbations for the split hemivariational inequality problem on Hadamard manifolds
• A trust region method for solving multicriteria optimization problems on Riemannian manifolds
• On the convergence of gradient descent for finding the Riemannian center of mass
• scientific article; zbMATH DE number 1785838 (Why is no real title available?)
• Subgradient algorithms on Riemannian manifolds of lower bounded curvatures
• Path-based incremental target level algorithm on Riemannian manifolds
• Gradient method for optimization on Riemannian manifolds with lower bounded curvature
• Fenchel conjugate via Busemann function on Hadamard manifolds
• Convergence of inexact steepest descent algorithm for multiobjective optimizations on Riemannian manifolds without curvature constraints
• Incremental subgradient algorithms with dynamic step sizes for separable convex optimizations
• On the relationship between the Kurdyka-Łojasiewicz property and error bounds on Hadamard manifolds
• A Grassmann manifold handbook: basic geometry and computational aspects
• Characterizations of the solution set of nonsmooth semi-infinite programming problems on Hadamard manifolds
• Optimality conditions and duality for multiobjective semi-infinite optimization problems with switching constraints on Hadamard manifolds
• Constraint qualifications and optimality conditions for nonsmooth multiobjective mathematical programming problems with vanishing constraints on Hadamard manifolds via convexificators
• Geodesic convexity of the symmetric eigenvalue problem and convergence of steepest descent
• The proximal point method for sum of DC functions in Hadamard manifolds
• Accelerated gradient methods on the Grassmann and Stiefel manifolds
• Rate of convergence analysis of proximal point algorithm on Riemannian manifold
• Iteration of complexity and asymptotic analysis of gradient-type methods with non-monotone line search on Riemannian manifolds

This page was built for publication: Convergence Analysis of Gradient Algorithms on Riemannian Manifolds without Curvature Constraints and Application to Riemannian Mass

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5148408)