Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Let $X$ be a Hilbert space with ON-basis $\lbrace e_n : ~ n \in \mathbb{N} \rbrace$. Furthermore let $A, ~ \Gamma : X \to X$ be linear operators with $A e_n = \alpha_n e_n$ and $\Gamma e_n = \gamma_n e_n$ respectively. Let $c,\lambda \in \mathbb{C}$.

My professor stated, that

$\left[ \begin{array}{ll} 0 & \Gamma \\ \lambda-A & c \Gamma \end{array} \right]$

is similar to the operator of multiplication by the sequence of matrices

$\left[ \begin{array}{ll} 0 & \gamma_n \\ \lambda-\alpha_n & c \gamma_n \end{array} \right] \in \mathbb{C}^{2 \times 2}$

Why is that correct? And what is even ment by "the operator of multiplication by the sequence of matrices"?

If you start calculating with some $v =(v_1, v_2) \in X \times X$ you get

$\left[ \begin{array}{ll} 0 & \Gamma \\ \lambda-A & c \Gamma \end{array} \right]v = \left[ \begin{array}{l} \Gamma v_2 \\ (\lambda -A) v_1 - c \Gamma v_1 \end{array} \right]$

and if you let $v_1 = \sum_{n \in \mathbb{N}} r_n e_n$, $v_2 = \sum_{n \in \mathbb{N}} s_n e_n$ this yields

$\left[ \begin{array}{ll} 0 & \Gamma \\ \lambda-A & c \Gamma \end{array} \right]v = \left[ \begin{array}{l} \sum_{n \in \mathbb{N}} s_n \gamma_n e_n \\ \lambda \sum_{n \in \mathbb{N}} r_n e_n - \sum_{n \in \mathbb{N}} r_n \alpha_n e_n - c \sum_{n \in \mathbb{N}} s_n \gamma_n e_n \end{array} \right]$

and this yields

$\left[ \begin{array}{ll} 0 & \Gamma \\ \lambda-A & c \Gamma \end{array} \right]v = \sum_{ n \in \mathbb{N}} \left[ \begin{array}{l} s_n \gamma_n e_n \\ \lambda r_n e_n - r_n \alpha_n e_n - c s_n \gamma_n e_n \end{array} \right]$.

But this doesn't help me at all. Any hint?

share|cite|improve this question
up vote 0 down vote accepted

My guess is that $A$ and $\Gamma$ share eigenvectors, $e_n$. The "big" matrix is written according to the product $X \times X$. (Bases of the two copies of $X$ need not be chosen here.) The small matrices (with $\alpha_n$ and $\gamma_n$) are only $2$-by-$2$ the way they are written, and they correspond to subspaces of $X \times X$ generated by $(e_n, 0)$ and $(0, e_n)$. They have to be extended to cover $X \times X$ by adding identity to properly act on $X \times X$. A formal definition may look like this:

Let $P_n:X\times X \to X\times X$ be defined as follows. Suppose $\left(u, v\right) = \left(\sum_{i=1}^\infty u_i e_i, \sum_{i=1}^\infty v_i e_i\right)$ and $(x, y) = \left(\sum_{i=1}^\infty x_i e_i, \sum_{i=1}^\infty y_i e_i\right) = P_n(u, v)$. Then \begin{align*} x_i & = \begin{cases} \gamma_i v_i & ; i = n\\ u_i & ; i \ne n \end{cases}\\ y_i & = \begin{cases} (\lambda - \alpha_i)u_i + c \gamma_i v_i & ; i = n \\ v_i & ; i \ne n \end{cases} \end{align*} These $P_n$ are operators defined by the small $2$-by-$2$ matrix you wrote above. The big operator should be the same as $\prod_{i=1}^\infty P_i$, which is at first glance informal but can be made formal by considering how it operates. Its action is exactly what you wrote.

share|cite|improve this answer
I edited the post according to what you wrote: indeed $A$ and $\Gamma$ share eigenvectors. Sorry I forgot that. I have to think a little about the rest of your post. – mjb Jan 28 '13 at 14:01

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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