Adaptive Hamiltonian variational integrators and applications to symplectic accelerated optimization
From MaRDI portal
Publication:5010240
Numerical optimization and variational techniques (65K10) Dynamical systems in optimization and economics (37N40) Numerical methods for Hamiltonian systems including symplectic integrators (65P10) Discretization methods and integrators (symplectic, variational, geometric, etc.) for dynamical systems (37M15) Canonical and symplectic transformations for problems in Hamiltonian and Lagrangian mechanics (70H15)
Abstract: It is well known that symplectic integrators lose their near energy preservation properties when variable step sizes are used. The most common approach to combine adaptive step sizes and symplectic integrators involves the Poincar'e transformation of the original Hamiltonian. In this article, we provide a framework for the construction of variational integrators using the Poincar'e transformation. Since the transformed Hamiltonian is typically degenerate, the use of Hamiltonian variational integrators based on Type II or Type III generating functions is required instead of the more traditional Lagrangian variational integrators based on Type I generating functions. Error analysis is provided and numerical tests based on the Taylor variational integrator approach of Schmitt, Shingel, Leok (2018) to time-adaptive variational integration of Kepler's 2-Body problem are presented. Finally, we use our adaptive framework together with the variational approach to accelerated optimization presented in Wibisono, Wilson, Jordan (2016) to design efficient variational and non-variational explicit integrators for symplectic accelerated optimization.
Recommendations
Cites work
- scientific article; zbMATH DE number 3850830 (Why is no real title available?)
- scientific article; zbMATH DE number 3790208 (Why is no real title available?)
- scientific article; zbMATH DE number 852353 (Why is no real title available?)
- A Software Package for the Numerical Integration of ODEs by Means of High-Order Taylor Methods
- A Stochastic Approximation Method
- A differential equation for modeling Nesterov's accelerated gradient method: theory and insights
- A variational perspective on accelerated methods in optimization
- Accelerating the cubic regularization of Newton's method on convex problems
- Adaptive subgradient methods for online learning and stochastic optimization
- An efficient overloaded method for computing derivatives of mathematical functions in MATLAB
- Automatic Hessians by reverse accumulation
- Backward Error Analysis for Numerical Integrators
- Directions for computing truncated multivariate Taylor series
- Discrete Hamiltonian variational integrators
- Discrete mechanics and variational integrators
- Discrete variational Hamiltonian mechanics
- Efficient recurrence relations for univariate and multivariate Taylor series coefficients
- Explicit adaptive symplectic (EASY) integrators: a scaling invariant generalisation of the Levi-Civita and KS regularisations
- Explicit adaptive symplectic integrators for solving Hamiltonian systems
- Explicit methods in extended phase space for inseparable Hamiltonian problems
- Exponential integrators
- Geometric Numerical Integration
- Geometric exponential integrators
- Hamiltonian Runge-Kutta type methods and their variational interpretation
- Hamiltonian dynamics on the symplectic extended phase space for autonomous and non-autonomous systems
- Introduction to Automatic Differentiation and MATLAB Object-Oriented Programming
- Introductory lectures on convex optimization. A basic course.
- Lagrangian and Hamiltonian Taylor variational integrators
- Lazy multivariate higher-order forward-mode AD
- Lie-Poisson Hamilton-Jacobi theory and Lie-Poisson integrators
- On Nonconvex Optimization for Machine Learning
- Performance of the Taylor series method for ODEs/DAEs
- Prolongation-collocation variational integrators
- Properties of Hamiltonian variational integrators
- Simulating Hamiltonian Dynamics
- Spectral variational integrators
- Stochastic discrete Hamiltonian variational integrators
- Stochastic variational integrators
- Symplectic integrators for long-term integrations in celestial mechanics
- Symplectic-energy-momentum preserving variational integrators
- The Development of Variable-Step Symplectic Integrators, with Application to the Two-Body Problem
- Time adaptive variational integrators: a space-time geodesic approach
- Time-step adaptivity in variational integrators with application to contact problems
- Variable time step integration with symplectic methods
Cited in
(17)- Adaptive Geometric Integrators for Hamiltonian Problems with Approximate Scale Invariance
- Practical perspectives on symplectic accelerated optimization
- On dissipative symplectic integration with applications to gradient-based optimization
- Exponential variational integrators using constant or adaptive time step
- Accelerated optimization on Riemannian manifolds via discrete constrained variational integrators
- Bregman dynamics, contact transformations and convex optimization
- Explicit adaptive symplectic integrators for solving Hamiltonian systems
- Multisymplectic Hamiltonian variational integrators
- Performance assessment of energy-preserving, adaptive time-step variational integrators
- Time-adaptive Lagrangian variational integrators for accelerated optimization
- A Herglotz-based integrator for nonholonomic mechanical systems
- Does variable step size ruin a symplectic integrator?
- Fast symplectic integrator for Nesterov-type acceleration method
- Time-step adaptivity in variational integrators with application to contact problems
- A variational formulation of accelerated optimization on Riemannian manifolds
- High-order symplectic Lie group methods on \(SO(n)\) using the polar decomposition
- Time adaptive variational integrators: a space-time geodesic approach
This page was built for publication: Adaptive Hamiltonian variational integrators and applications to symplectic accelerated optimization
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5010240)