MathOverflow is a question and answer site for professional mathematicians. 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

In comparing the norm of two operators, I come across the following problem.

Let $S\in M_{n}(\mathbb{R})$ be a symmetric matrix. $D_1=diag(\alpha_1,\cdots,\alpha_n)$, $D_2=diag(\beta_1,\cdots,\beta_n)$, with $\alpha_1\ge\cdots\ge\alpha_n\ge0, ~\beta_1\ge\cdots\ge\beta_n\ge 0$. Is it true that $Tr[(D_1S^2D_1D_2+I)^{-1}]\le Tr[(SD_1^2SD_2+I)^{-1}]$?

Updated The above statment is not true as shown by Gerald Edgar. Now with the same notation, define $A=D_2SD_1^2SD_2, B=D_2D_1S^2D_1D_2$ and $$X_{k+1}=\frac{1}{2}(X_k+AX_k^{-1}),~~ X_0=I.$$ $$Y_{k+1}=\frac{1}{2}(Y_k+BY_k^{-1}), ~~Y_0=I.$$

Is it true $Tr X_{k}\le Tr Y_{k}$ for all $k\ge 1$?

When $k=1$, it is true.

share|cite|improve this question
There surely is some motivation behind all this? – Mariano Suárez-Alvarez Apr 25 '10 at 15:48
My purpose is to compare norms of two operators, it has been done. The updated one I created is more than what I want, I don't know how it occurs to me, but it looks interesting, so I post it. – Sunni Apr 25 '10 at 17:16
up vote 8 down vote accepted

I just wrote down some random 2x2 example, and it had the wrong sense for the inequality. A = [2,1;1,0], D1=D2=diag(2,1) .

share|cite|improve this answer

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.