Newton-Based Mixed Precision Iterative Refinement for Large-Scale Sparse Continuous-Time Algebraic Riccati Equations

DOI10.48550/arXiv.2607.14742MaRDI QIDQ7284093arXivFDO

Authors Juan Zhang, Yang Zhou

Publication date 16 July 2026

Published in arXiv

Tldr We propose Newton-based iterative refinement for solving large-scale sparse continuous-time algebraic Riccati equations in lower precision with high-precision residual evaluation and updates. We provide a rounding error analysis that relates stable refinement to a Lyapunov operator conditioning threshold, and demonstrate speedup on both dense and sparse systems.

Full work available at URL https://doi.org/10.48550/arXiv.2607.14742

Mathematics Subject Classification ID

Numerical methods for matrix equations (65F45) Computational methods for sparse matrices (65F50) Roundoff error (65G50) Complexity and performance of numerical algorithms (65Y20)

arXiv classification

math.NA

Cites work

• A class of Petrov-Galerkin Krylov methods for algebraic Riccati equations
• A new set of invariants for linear systems--Application to reduced order compensator design
• A Schur method for solving algebraic Riccati equations
• A structure-preserving doubling algorithm for continuous-time algebraic Riccati equations
• Accelerating the solution of linear systems by iterative refinement in three precisions
• An inexact low-rank Newton-ADI method for large-scale algebraic Riccati equations
• Approximation of Large-Scale Dynamical Systems
• Linear model reduction and solution of the algebraic Riccati equation by use of the sign function†
• Low-rank generalized alternating direction implicit iteration method for solving matrix equations
• Mixed precision algorithms in numerical linear algebra
• Mixed-precision algorithms for solving the Sylvester matrix equation
• Mixed-precision iterative refinement for low-rank Lyapunov equations
• RADI: a low-rank ADI-type algorithm for large scale algebraic Riccati equations
• The projected Newton-Kleinman method for the algebraic Riccati equation
• The University of Florida sparse matrix collection
• Three-precision GMRES-based iterative refinement for least squares problems
• Truncated QR factorization with pivoting in mixed precision
• Using the Matrix Sign Function to Compute Invariant Subspaces

This page was built for publication: Newton-Based Mixed Precision Iterative Refinement for Large-Scale Sparse Continuous-Time Algebraic Riccati Equations

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7284093)