6
$\begingroup$

When I compute the eigenvectors of a certain matrix, the first of them is composed entirely of a single constant.

What properties of a matrix lead to this result?

Update

By "a vector composed entirely of a constant" I mean n repetitions of a constant comprising a length-n vector.

$\endgroup$
  • $\begingroup$ What do you mean by a vector being "composed entirely of a single constant"? If possible, please provide a concrete example. $\endgroup$ – hmakholm left over Monica Oct 8 '11 at 0:10
  • 2
    $\begingroup$ Each row of the matrix has the same sum. $\endgroup$ – André Nicolas Oct 8 '11 at 0:11
  • $\begingroup$ My guess is it means something like $(3,3,3,3,3,3)$; every component is the same. $\endgroup$ – Michael Hardy Oct 8 '11 at 0:22
  • 1
    $\begingroup$ Special case of this phenomenon are stochastic matrices, which are of importance in several areas of mathematics. $\endgroup$ – Martin Sleziak Oct 8 '11 at 8:29
7
$\begingroup$

Take an $n \times n$ matrix $A$, and suppose that $v$ is an eigenvector of $A$, with all entries of $v$ equal to a constant $k$. Naturally, $k \ne 0$. Let $\lambda$ be the eigenvalue of $A$ that has $v$ as an eigenvector. If $(b_1, b_2, \dots, b_n)$ is any row of $A$, then by the definition of eigenvalue and eigenvector, we have $$kb_1+kb_2+\cdots +kb_n=\lambda k,$$ from which we conclude that $b_1+b_2+\cdots+b_n=\lambda$. It follows that each row sum of the matrix is equal to $\lambda$.

Conversely, suppose that all row sums of $A$ are equal to $\sigma$. Let $v$ be the vector with all entries equal to $1$. Then $Av$ is a vector with all entries equal to $\sigma$, which means that $v$ is an eigenvector of $A$ with eigenvalue $\sigma$.

Thus $A$ has an eigenvector with all entries equal if and only if all row sums of $A$ are equal.

| cite | improve this answer | |
$\endgroup$
2
$\begingroup$

Exercise: a matrix has an eigenvector with all its entries equal if and only if all rows of the matrix have the same sum.

| cite | improve this answer | |
$\endgroup$

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.