Please help me calculate eigenvalues and eigenvectors for a square matrix, with $n$ rows and $n$ columns, and $-1$ in each cell:$$\begin{bmatrix} -1 & \,\,\cdots & -1\\\,\,\vdots \strut& \,\,\ddots & \,\,\vdots \strut\\ -1 &\,\,\cdots & -1\end{bmatrix}$$

I guess that this is a $0$ and $n$, vectors $(1,1,\ldots,1)$ and $(1,-1,0,\ldots)$, $(1,0,-1,0,\ldots),\ldots$

How to prove it?

  • $\begingroup$ To be clear, you mean a matrix of the form $$\begin{bmatrix} -1 & \cdots & -1\\\,\vdots \strut& \,\ddots & \,\vdots \strut\\ -1 &\cdots & -1\end{bmatrix}\qquad ?$$ $\endgroup$ – Zev Chonoles Feb 28 '13 at 20:00
  • $\begingroup$ Yes, that's correct matrix $\endgroup$ – Steve Feb 28 '13 at 20:03

For $i=1\ldots n$ let $e_i$ be the $i$-th vector of the canonical basis of $\mathbb R^n$, i.e. $$ (\vec e_i)_j = \begin{cases} 1 & \text{if }i=j \\ 0 & \text{if }i\neq j\end{cases} $$ If $M_{ij}=-1$ $\forall i,j$ then, as you say, $\vec v_i=\vec e_1-\vec e_i$ ($i=2\ldots n$) are all eigenvectors of $M$ with eigenvalue $0$. In fact for all $j=1\ldots n$ $$ (M\vec v_i)_j = \sum_{\ell=1}^n M_{j\ell}(\vec v_i)_\ell = (-1)\cdot (\vec v_i)_1 + (-1)\cdot(\vec v_i)_i = -1 -(-1) = 0 $$ Therefore $M\vec v_i = \vec 0 = 0\cdot\vec v_i$.

A similar computation shows that $\vec u=(1\ldots 1)$ is an eigenvector of $M$ with eigenvalue $1$.

$\vec u,\vec v_2\ldots \vec v_n$ are $n$ linearly independent eigenvectors, i.e. they form a basis of $\mathbb R^n$ so no more linearly independent eigenvectors exist.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.