The condition number of a randomly perturbed matrix

Abstract

Let $M$ be an arbitrary $n$ by $n$ matrix. We study the condition number a random perturbation $M+N_n$ of $M$, where $N_n$ is a random matrix. It is shown that, under very general conditions on $M$ and $M_n$, the condition number of $M+N_n$ is polynomial in $n$ with very high probability. The main novelty here is that we allow $N_n$ to have discrete distribution.

8 pages, no figures, to appear, STOC '07

References in corpus (2)

• Smoothed analysis of algorithms
• Invertibility of random matrices: norm of the inverse