0
$\begingroup$

I know that with Sherman–Morrison formula, we can simply calculate $(\mathbf{I}+\mathbf{v}\mathbf{v}^H)^{-1}$, where $\mathbf{v}$ is a column vetor. What about the case of the summation, i.e., $(\mathbf{I}+\sum_{i}\mathbf{v}_i\mathbf{v}_i^H)^{-1}$

$\endgroup$
1
0
$\begingroup$

In general $Z=\sum_{i=1}^r u_iv_i^*$ has rank $r$. When $r/n$ is small, using the Woodbury formula has a low complexity. This requires finding $U,C,V$ s.t. $Z=UCV$ where $U\in M_{n,r},C\in GL_r,V\in M_{r,n}$.

I do not think that the hypothesis $u_i=v_i$ greatly reduces the complexity.

$\endgroup$

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.