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


I have a matrix : $W=I+U^TV$

  • $dim(W)=(D,D)$
  • $dim(U)=dim(V)=(N,D)$ with $N < < D$

I need it to be orthogonal ie $W^TW=I$

which gives me : $V^TU+U^TV+V^TUU^TV=0$

From that point, i don't know where to go. Have anyone got some ideas about that issue ?



share|cite|improve this question
I assume your second term should be $U^T V$. – Jeff Schenker Feb 25 '11 at 17:14
Yes, i corrected it – GuillaumeThomas Feb 25 '11 at 17:26

Do you want a general solution or simply a good family of examples? If you want examples, let $V=-U=\sqrt{2} S^T$ where $S$ is an isometry from $\mathbb{R}^N$ into $\mathbb{R}^D$.

share|cite|improve this answer

This should be a comment. I think this question is not apt here. You should consider posting questions like these at math.stackexchange.

Given $U$ and if it is full rank, we can choose $$V = -2 \left(UU^T \right)^{-1}U$$ and the matrix $$W = I - 2 U^T \left(UU^T \right)^{-1}U$$ is an orthonormal matrix. A special case of this is provided by Jeff.

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.