Convergence and cycling in Walker-type saddle search algorithms
From MaRDI portal
Publication:5359495
DOI10.1137/16M1087199zbMATH Open1375.65088arXiv1607.08875MaRDI QIDQ5359495FDOQ5359495
Authors: Antoine Levitt, C. Ortner 
Publication date: 25 September 2017
Published in: SIAM Journal on Numerical Analysis (Search for Journal in Brave)
Abstract: Algorithms for computing local minima of smooth objective functions enjoy a mature theory as well as robust and efficient implementations. By comparison, the theory and practice of saddle search is destitute. In this paper we present results for idealized versions of the dimer and gentlest ascent (GAD) saddle search algorithms that show-case the limitations of what is theoretically achievable within the current class of saddle search algorithms: (1) we present an improved estimate on the region of attraction of saddles; and (2) we construct quasi-periodic solutions which indicate that it is impossible to obtain globally convergent variants of dimer and GAD type algorithms.
Full work available at URL: https://arxiv.org/abs/1607.08875
Recommendations
- A dimer-type saddle search algorithm with preconditioning and linesearch
- Global optimization-based dimer method for finding saddle points
- An iterative minimization formulation for saddle point search
- Shrinking dimer dynamics and its applications to saddle point search
- A local minimax method using virtual geometric objects. I. For finding saddles
Numerical optimization and variational techniques (65K10) Existence theories for optimal control problems involving partial differential equations (49J20)
Cites Work
- Numerical Optimization
- Trust Region Methods
- Minimum action method for the study of rare events
- Shrinking dimer dynamics and its applications to saddle point search
- The string method as a dynamical system
- The gentlest ascent dynamics
- An iterative minimization formulation for saddle point search
- A dimer-type saddle search algorithm with preconditioning and linesearch
- On the global convergence of path-following methods to determine all solutions to a system of nonlinear equations
- Optimization-based shrinking dimer method for finding transition states
Cited In (19)
- A constrained gentlest ascent dynamics and its applications to finding excited states of Bose-Einstein condensates
- Convergence analysis of discrete high-index saddle dynamics
- Efficient sampling of saddle points with the minimum-mode following method
- An iterative minimization formulation for saddle point search
- Title not available (Why is that?)
- Discretization and index-robust error analysis for constrained high-index saddle dynamics on the high-dimensional sphere
- A dimer-type saddle search algorithm with preconditioning and linesearch
- Error estimates for Euler discretization of high-index saddle dynamics
- Global optimization-based dimer method for finding saddle points
- Error estimate for semi-implicit method of sphere-constrained high-index saddle dynamics
- Variational Phase Field Formulations of Polarization and Phase Transition in Ferroelectric Thin Films
- Simplified gentlest ascent dynamics for saddle points in non-gradient systems
- Using Witten Laplacians to Locate Index-1 Saddle Points
- Switched diffusion processes for non-convex optimization and saddle points search
- Constrained shrinking dimer dynamics for saddle point search with constraints
- Probabilistic error estimate for numerical discretization of high-index saddle dynamics with inaccurate models
- Staying the course: iteratively locating equilibria of dynamical systems on Riemannian manifolds defined by point-clouds
- Enclosure of all index-1 saddle points of general nonlinear functions
- Shrinking dimer dynamics and its applications to saddle point search
This page was built for publication: Convergence and cycling in Walker-type saddle search algorithms
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5359495)
