AI & ML
Variational integrators for time-dependent Lagrangian systems on Lie groups
arXiv cs.NA0 views
arXiv:2611.25313v1 Announce Type: new Abstract: We develop fixed-step variational integrators for time-dependent Lagrangian systems on Lie groups, derive their reduced dynamics, and analyze their long-time behavior by backward error analysis. We derive the continuous Euler-Lagrange equations, energy balance, and Noether conservation laws, and construct time-indexed discrete Lagrangians using retractions, with the midpoint-Cayley scheme treated explicitly. On their regularity domains, the resulting phase-space maps are symplectic. For left-invariant discrete Lagrangians, they reduce to Lie-Poisson maps that preserve coadjoint orbits and, when $G$ is connected, every Casimir function. A variational error analysis establishes second-order accuracy of the midpoint-Cayley method. We also construct an exactly symplectic, projectable extension on $T^*(\mathbb R\times G)$ whose formal modified Hamiltonian is affine in the momentum conjugate to time and admits symmetry reduction. Under uniform analyticity, coefficient bounds, and domain-containment assumptions, optimal truncation yields exponentially accurate interpolation and near-conservation of the modified extended Hamiltonian over exponentially long intervals. For a connected attitude network on $SO(3)^N$ with damping and uniformly positive, non-increasing coupling weights, we prove local asymptotic stability of the static consensus set for the continuous dynamics and exact preservation of total spatial momentum by the discrete scheme. Numerical experiments on the network and on a time-dependent rigid body on $SO(3)$ illustrate second-order convergence and the extended energy balance.
Source:https://arxiv.org/abs/2611.25313PreviousA fast iterative solver for lopsided Sylvester matrix equationsNextA GPU Interior-Point Solver with Macro-Element Condensation for Equilibrium-Based Plasticity
Related posts
VESTA: Visual Exploration with Statistical Tool Agents
0arXiv stat.CO
An interpretable semiparametric Bayesian machine learning model for high-dimensional circulating epigenetic signatures of multiple myeloma
0arXiv stat.CO
Quantum Nondecimated Wavelet Transform: Theory, Circuits, and Applications
0arXiv stat.CO
dapper: Data Augmentation for Private Posterior Estimation in R
0arXiv stat.CO
HAVAPlot: A Web-Based Scientific Visualization and Post-Processing Platform
0arXiv stat.CO
ISC Stormcast For Friday, November 20th, 2026 https://isc.sans.edu/podcastdetail/10188, (Fri, Nov 20th)
0SANS ISC