Convergence of the vertical gradient flow for the Gaussian Monge problem
From MaRDI portal
Publication:6201384
DOI10.3934/JCD.2023008arXiv2305.19788OpenAlexW4388468549MaRDI QIDQ6201384FDOQ6201384
Authors: Erik Jansson, Klas Modin 
Publication date: 20 February 2024
Published in: Journal of Computational Dynamics (Search for Journal in Brave)
Abstract: We investigate a matrix dynamical system related to optimal mass transport in the linear category, namely, the problem of finding an optimal invertible matrix by which two covariance matrices are congruent. We first review the differential geometric structure of the problem in terms of a principal fiber bundle. The dynamical system is a gradient flow restricted to the fibers of the bundle. We prove global existence of solutions to the flow, with convergence to the polar decomposition of the matrix given as initial data. The convergence is illustrated in a numerical example.
Full work available at URL: https://arxiv.org/abs/2305.19788
Optimal transportation (49Q22) Factorization of matrices (15A23) Approximation methods and numerical treatment of dynamical systems (37M99)
Cites Work
- Algorithm 432 [C2]: Solution of the matrix equation AX + XB = C [F4]
- Polar factorization and monotone rearrangement of vector‐valued functions
- The Toda lattice. II. Existence of integrals
- Computing the Polar Decomposition—with Applications
- Title not available (Why is that?)
- Dynamical systems that sort lists, diagonalize matrices, and solve linear programming problems
- On the Separation of Two Matrices
- Information geometry and its applications
- Geometry of matrix decompositions seen through optimal transport and information geometry
- Shape analysis via gradient flows on diffeomorphism groups
This page was built for publication: Convergence of the vertical gradient flow for the Gaussian Monge problem
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6201384)
