Comparing the methods of alternating and simultaneous projections for two subspaces

DOI10.1016/J.LAA.2023.12.007MaRDI QIDQ6146685zbMATH OpenOpenAlexFDO

Authors Simeon Reich, Rafał Zalas

Publication date 15 January 2024

Published in Linear Algebra and its Applications (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/2306.12219

zbMATH Keywords

Friedrichs anglemethod of alternating projectionsmethod of simultaneous projectionsprincipal anglesrates of convergence

Mathematics Subject Classification ID

Rate of convergence, degree of approximation (41A25) Simultaneous approximation (41A28) Abstract approximation theory (approximation in normed linear spaces and other abstract spaces) (41A65)

Abstract: We study the well-known methods of alternating and simultaneous projections when applied to two nonorthogonal linear subspaces of a real Euclidean space. Assuming that both of the methods have a common starting point chosen from either one of the subspaces, we show that the method of alternating projections converges significantly faster than the method of simultaneous projections. On the other hand, we provide examples of subspaces and starting points, where the method of simultaneous projections outperforms the method of alternating projections.

Review text

The best approximation problem for two linear subspaces \(A\) and \(B\) of \(\mathbb R^d\) is to find a point in the intersection \(A \cap B\) which is the closest to any given starting point \(x \in \mathbb R^d\). In this paper, the authors investigate two well known methods for solving this problem, the method of alternating projections and the method of simultaneous projections. The main results are to provide a set of conditions under which the former converges faster than the latter and to provide a different set of conditions under which the opposite behaviour can be observed. The Friedrichs angle between the two subspaces \(A\) and \(B\) plays an important role, but it is not sufficient to provide a complete characterization. To this end, the authors utilize the (more general) concept of principal angles.

Recommendations

• Acceleration schemes for the method of alternating projections
• Accelerating the convergence of the method of alternating projections
• Optimal rates of linear convergence of the averaged alternating modified reflections method for two subspaces
• Alternating Projections on Manifolds
• On angles, projections and iterations

Cites work

• A gentle guide to the basics of two projections theory
• A limit theorem for projections
• Accelerating the convergence of the method of alternating projections
• Alternating projections, remotest projections, and greedy approximation
• Angles between infinite dimensional subspaces with applications to the Rayleigh-Ritz and alternating projectors methods
• Arbitrarily slow convergence of sequences of linear operators
• Best approximation in inner product spaces
• Characterizing arbitrarily slow convergence in the method of alternating projections
• Circumcentering the Douglas-Rachford method
• Convex analysis and monotone operator theory in Hilbert spaces
• Directional asymptotics of Fejér monotone sequences
• Error bounds for the method of alternating projections
• Error bounds for the method of simultaneous projections with infinitely many subspaces
• Generalization of the Trotter-Lie formula
• scientific article; zbMATH DE number 5835628 (Why is no real title available?)
• scientific article; zbMATH DE number 1009689 (Why is no real title available?)
• scientific article; zbMATH DE number 1460605 (Why is no real title available?)
• scientific article; zbMATH DE number 2148414 (Why is no real title available?)
• scientific article; zbMATH DE number 3451986 (Why is no real title available?)
• scientific article; zbMATH DE number 2111135 (Why is no real title available?)
• scientific article; zbMATH DE number 3230744 (Why is no real title available?)
• Iterative methods for fixed point problems in Hilbert spaces
• Jordan's principal angles in complex vector spaces
• Linear integral equations
• Necessary conditions for linear convergence of iterated expansive, set-valued mappings
• Numerical Methods for Computing Angles Between Linear Subspaces
• On angles, projections and iterations
• On operator ranges
• On Projection Algorithms for Solving Convex Feasibility Problems
• On rings of operators. Reduction theory
• On the spectra of sums of orthogonal projections with applications to parallel computing
• Optimal rates of linear convergence of relaxed alternating projections and generalized Douglas-Rachford methods for two subspaces
• Optimal rates of linear convergence of the averaged alternating modified reflections method for two subspaces
• Polynomial estimates for the method of cyclic projections in Hilbert spaces
• Projection methods: an annotated bibliography of books and reviews
• Subspaces, angles and pairs of orthogonal projections
• The optimal error bound for the method of simultaneous projections
• The rate of linear convergence of the Douglas-Rachford algorithm for subspaces is the cosine of the Friedrichs angle

This page was built for publication: Comparing the methods of alternating and simultaneous projections for two subspaces

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6146685)