← Tags

Condition Number

Mathematics Numerical computing

[Linear algebra] Why does the condition number describe numerical difficulty? (Rewritten from an earlier blog post)

Explain how a matrix condition number affects the accuracy of a linear-system solution. Keep the matrix fixed, perturb the right-hand side, and derive a bound on relative error amplification. Use Python to examine singular values, ellipses, error directions, and inaccurate solutions with small residuals, while distinguishing problem conditioning from algorithm stability.

Mathematics Numerical computing

[Linear algebra] Why are Hilbert matrices difficult for computers? (Rewritten from an earlier blog post)

A Hilbert matrix has an inverse, yet small input errors can greatly change the solution of a linear system. Use an integral to prove invertibility and the condition number to explain error amplification. Compare residuals and solution errors in Python, and plot the difference between rebuilding inputs at high precision and reusing inputs rounded to double precision.