2
$\begingroup$

By the circulant matrix $C$ in $M_n(\mathbb{R})$, we mean that $$C=[e_n|e_1|\cdots|e_{n-1}]$$ where $e_1,\cdots,e_n$ are the standard basis vectors in $\mathbb{R}^n$. It is well-known that
$$C=\mathbf{F}\operatorname{Diag}(1,w,\cdots,w^{n-1})\mathbf{F}^{-1}$$ where $\mathbf{F}$ is just the discrete Fourier matrix and $w=\exp(\frac{2\pi i}{n})$

Let $U=(u_{ij})$ be a upper triangular matrix. We say it is supper upper triangular if $u_{i, i+1}=0$.

Let $U$ be a supper upper triangular whose entries are either 0 or 1. Is $C+U$ diagonalizable?

$\endgroup$
6
  • 1
    $\begingroup$ That circulant matrix is also a permutation matrix $\endgroup$ Commented May 29, 2023 at 9:48
  • $\begingroup$ @Rodrigo de Azevedo, Yes. If it would be valid, one may think about the extension on a larger category of matrices containing the circulant matrix. $\endgroup$
    – ABB
    Commented May 29, 2023 at 9:52
  • $\begingroup$ Since $\bf F$ is unitary, its inverse is easy to compute $\endgroup$ Commented May 29, 2023 at 9:54
  • $\begingroup$ Yes, the inverse is just the transpose. $\endgroup$
    – ABB
    Commented May 29, 2023 at 9:57
  • 2
    $\begingroup$ You mean, upper triangular with zero diagonal? $\endgroup$ Commented May 30, 2023 at 12:18

1 Answer 1

2
$\begingroup$

Here's a counterexample: For $U$ take $\left(\begin{smallmatrix} 0& 0 &0& 0& 1& 1& 1& 1\\ 0& 0 &0 &1& 0& 1& 1& 1\\ 0& 0 &0& 0& 1& 0& 0& 1\\ 0& 0 &0& 0& 0& 0& 0& 1\\ 0& 0 &0& 0& 0& 0& 1& 0\\ 0& 0 &0& 0& 0& 0& 0& 0\\ 0& 0 &0& 0& 0& 0& 0& 0\\ 0& 0 &0& 0& 0& 0& 0& 0 \end{smallmatrix}\right)$.

Then the rational canonical form of $C+U$ is $\left(\begin{smallmatrix} -1 & 1 & 0 & 0 & 0 & 0 & 0 & 0\\ 0 &-1& 0 & 0 & 0 & 0 & 0 & 0\\ 0 &0 & 0 & 1 & 0 & 0 & 0 & 0\\ 0 &0 & 0 & 0 & 1 & 0 & 0 & 0\\ 0 &0 & 0 & 0 & 0 & 1 & 0 & 0\\ 0 &0 & 0 & 0 & 0 & 0 & 1 & 0\\ 0 &0 & 0 & 0 & 0 & 0 & 0 & 1\\ 0 &0 & 1 & 1 & -1 & 4& -2 & 2 \end{smallmatrix}\right)$

$\endgroup$
5
  • $\begingroup$ If you want to check this, here's some Magma code: Q:=Rationals(); M:=MatrixAlgebra(Q,8); U:=M![ 0, 0, 0, 0, 1, 1, 1, 1, 0, 0, 0, 1, 0, 1, 1, 1, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]; C:=M![ 0,1,0,0,0,0,0,0, 0,0,1,0,0,0,0,0, 0,0,0,1,0,0,0,0, 0,0,0,0,1,0,0,0, 0,0,0,0,0,1,0,0, 0,0,0,0,0,0,1,0, 0,0,0,0,0,0,0,1, 1,0,0,0,0,0,0,0]; JordanForm(C+U); $\endgroup$ Commented May 30, 2023 at 12:51
  • $\begingroup$ Thanks, it is so helpful. $\endgroup$
    – ABB
    Commented May 30, 2023 at 15:00
  • $\begingroup$ If you want to know how I found this, it was actually a rather stupid method. I tried Random(0,1) examples on Magma, starting with $n=4$ and gradually increasing until an example appeared. Nothing clever at all. $\endgroup$ Commented May 30, 2023 at 15:02
  • $\begingroup$ Actually, your suggestion made me move forward in a research project, related to graph signal processing! $\endgroup$
    – ABB
    Commented May 30, 2023 at 16:17
  • $\begingroup$ @ABB : ----- Nice! $\endgroup$ Commented May 30, 2023 at 16:59

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .