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

This question concerns a matrix $A\in\mathbb{C}^{n\times n}$, whose inverse we know the action of on a particular vector $v\in\mathbb{C}^{n}$. If we know that $A^{-1}v = u$, is there any way to express $(A+\alpha I)^{-1}v$? We have that $\alpha \in \mathbb{C}$ but it is not bounded in magnitude and $A$ is any matrix.

share|cite|improve this question

The only case in which you have enough information is if $u$ is a scalar multiple of $v$. If $u = t v$ with $t \ne 0$, i.e. $A v = v/t$, then $(A + \alpha I) v = (1/t + \alpha) v$ so either $(A + \alpha I)^{-1}$ doesn't exist or $(A + \alpha I)^{-1} v = (1/t + \alpha)^{-1} v$. If $u$ is not a scalar multiple of $v$, you don't have enough information to determine $(A + \alpha I)^{-1} v$.

share|cite|improve this answer
Also known as "eigenvector"? – Federico Poloni Aug 5 '11 at 7:13

By adding a constant identity to $A$, you increase all of its eigenvalues by $\alpha$. I rewrite your transformation as $$(A+\alpha I)^{-1}v = X(\Lambda+aI)^{-1}X^{-1}v,$$ where $X$ is a set of eigenvectors for $A$ and $\Lambda$ is the eigevalues diagonal. We have that $u = X\Lambda^{-1}X^{-1}v$, or $$X\Lambda X^{-1}u = v.$$ We can obtain $$(A+\alpha I)^{-1}v=X(\Lambda+aI)^{-1}\Lambda X^{-1}u=X(I-a(\Lambda+aI)^{-1}) X^{-1}u=u-aX(\Lambda+aI)^{-1}X^{-1}u$$ due to $\frac{\lambda_i}{\lambda_i+\alpha}=1-\frac{\alpha}{\lambda_i+\alpha}$.

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.