Linear interval equations: Midpoint preconditioning may produce a 100\% overestimation for arbitrarily narrow data even in case n = 4

DOI10.1007/S11155-005-3033-5MaRDI QIDQ2484034zbMATH OpenOpenAlexFDO

Authors Jiří Rohn

Publication date 2 August 2005

Published in Reliable Computing (Search for Journal in Brave)

Full work available at URL https://doi.org/10.1007/s11155-005-3033-5

zbMATH Keywords

interval hulllinear interval equationsmidpoint preconditioningnumerical exampleoverestimation

Mathematics Subject Classification ID

Linear equations (linear algebraic aspects) (15A06) Numerical computation of matrix norms, conditioning, scaling (65F35) Interval and finite arithmetic (65G30)

Review text

The author discusses the solution of linear interval systems (that is, computation of the interval hull of the solution set) by means of so-called midpoint preconditioning techniques, which usually produces good results for regular matrices with small spectral radii. However, the author shows that preconditioning is not an universal remedy. He presents an example of a \(4\times 4\) interval system, such that midpoint preconditioning produces a \(100\%\) overestimation for arbitrary narrow data, what can be considered as an utterly false result.

Recommendations

• On overestimation produced by the interval Gaussian algorithm
• The hull of preconditioned interval linear equations
• New operator and method for solving real preconditioned interval linear equations
• Enclosing the solution set of overdetermined systems of interval linear equations
• Bounding the Solution of Interval Linear Equations

Cites work

• A Comparison of some Methods for Solving Linear Interval Equations
• A simple derivation of the Hansen-Bliek-Rohn-Ning-Kearfott enclosure for linear interval equations
• Bounding the Solution of Interval Linear Equations
• Compatibility of approximate solution of linear equations with given error bounds for coefficients and right-hand sides
• Computing Exact Componentwise Bounds on Solutions of Lineary Systems with Interval Data is NP-Hard
• Computing the norm ∥A∥∞,1 is NP-hard^∗
• scientific article; zbMATH DE number 804623 (Why is no real title available?)
• Systems of linear interval equations

Cited in

(3)

• On overestimation produced by the interval Gaussian algorithm
• The hull of preconditioned interval linear equations
• Overestimation in Linear Interval Equations

Describes a project that uses

Uses Software

• Matlab

This page was built for publication: Linear interval equations: Midpoint preconditioning may produce a 100\% overestimation for arbitrarily narrow data even in case \(n = 4\)

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2484034)