From: magya_bloom on
anyone knows the above (that is relative to the trace(AB) where A and
B are Hermitian)? thanks.
From: Maarten Bergvelt on
On 2010-07-22, magya_bloom(a)yahoo.com <magya_bloom(a)yahoo.com> wrote:
> anyone knows the above (that is relative to the trace(AB) where A and
> B are Hermitian)? thanks.

Why don't you choose your favorite basis of that space and do the
Gram-Schmidt song-and-dance routine?

--
Maarten Bergvelt
From: Gerry on
On Jul 23, 6:57 am, Maarten Bergvelt <be...(a)math.uiuc.edu> wrote:
> On 2010-07-22, magya_bl...(a)yahoo.com <magya_bl...(a)yahoo.com> wrote:
>
> > anyone knows the above (that is relative to the trace(AB) where A and
> > B are Hermitian)? thanks.
>
> Why don't you choose your favorite basis of that space and do the
> Gram-Schmidt song-and-dance routine?

Maybe OP wants a formula that works for all n.
--
GM
From: Stephen Montgomery-Smith on
magya_bloom(a)yahoo.com wrote:
> anyone knows the above (that is relative to the trace(AB) where A and
> B are Hermitian)? thanks.

Matrices which are diagonal with diagonal entries (0,0,...,0,1,0,...0),
and matrices which for i not equal to j have the entries a_ij and a_ji
equal to 1/sqrt(2) and all other entries zero.
From: Stephen Montgomery-Smith Montgomery-Smith on
On Jul 22, 8:25 pm, Stephen Montgomery-Smith
<step...(a)math.missouri.edu> wrote:
> magya_bl...(a)yahoo.com wrote:
> > anyone knows the above (that is relative to the trace(AB) where A and
> > B are Hermitian)? thanks.
>
> Matrices which are diagonal with diagonal entries (0,0,...,0,1,0,...0),
> and matrices which for i not equal to j have the entries a_ij and a_ji
> equal to 1/sqrt(2) and all other entries zero.

I forgot the matrices in which a_ij is i/sqrt(2) and a_ji is -i/
sqrt(2).