Could you guys give me some hints on this homework?

Find inverse of $\mathbf{I} + \mathbf{ab}^\intercal$. Hint: try to form $c\mathbf{I} + d\mathbf{ab}^\intercal$ and solve for $c,d$. What happens when $\mathbf{b}^\intercal \mathbf{a} = -1$?


Not quite sure what the hint is trying to say. Perhaps Inverse is of the form $c\mathbf{I}+d\mathbf{ab}^\intercal$.

$\begin{align*} (\mathbf{I}+\mathbf{ab}^\intercal)(c\mathbf{I}+d\mathbf{ab}^\intercal) &= \mathbf{I}\\ c\mathbf{I}+(c+d)\mathbf{ab}^\intercal + d(\mathbf{ab}^\intercal)(\mathbf{ab}^\intercal) &= \mathbf{I} \end{align*}$

if $\mathbf{b}^\intercal\mathbf{a} = -1$ then

$\begin{align*} c\mathbf{I} + c\mathbf{ab}^\intercal &= \mathbf{I} \end{align*}$

This doesn't look like I am going down the right path.


First notice that $b^Ta$ is a scalar so $$ab^Tab^T=(b^Ta)ab^T$$ and then we find

$$cI+(c+d(1+b^Ta))ab^T=I$$ hence in the case $1+b^Ta\ne0$ we choose $c=1$ and $d=\frac{-1}{1+b^Ta}$ to find the desired inverse.

In the case $b^Ta=-1$ then $$(I+ab^T)a=a-a=0$$ so $a$ is an eigenvector associated to the eigenvalue $0$ hence $I+ab^T$ isn't invertible.

  • $\begingroup$ ah thanks, I guess I was on the right track $\endgroup$
    – bdeonovic
    Sep 6 '14 at 16:45
  • $\begingroup$ Yes you were! You're welcome. $\endgroup$
    – user63181
    Sep 6 '14 at 16:48

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.