Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. 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

Is it possible to define a Taylor expansion for matrices ? Can I use functional derivative ?

More precisely I have to calculate something like : $\ln(A+B)$ using a Taylor expansion, where $A$ and $B$ are hermitian matrices which depend also on $x\in \mathbb{R}^3$. My idea is to use functional derivative but I don't know if the result is correct!

share|cite|improve this question
up vote 4 down vote accepted

In one particular case there is very good approach. Denote $C=A+B-I$, if $\operatorname{Sp}(C)\subset\mathbb{D}$ then we can use Tailor expansion for the $$ \ln(I+C)=\sum\limits_{n=1}^\infty\frac{(-1)^{n+1}}{n}C^n=\sum\limits_{n=1}^\infty\frac{(-1)^{n+1}}{n}(A+B-I)^n $$ If this conditions are not satisfied you should apply general theory - holomorphic functional calculus. In order to $\ln(A+B)$ make sense it is necessary that $\operatorname{Sp}(A+B)\subset\mathbb{C}^\times$, because $\ln\in\mathcal{O}(\mathbb{C}^\times)$. In this case we take arbitrary contour $\gamma$ containing $\operatorname{Sp}(A+B)$ and "simply" calculate $$ \ln(A+B)=\frac{1}{2i\pi}\int\limits_\gamma\ln(\lambda)(\lambda I-A-B)^{-1}d\lambda $$

share|cite|improve this answer
More explicitly: the series for the matrix logarithm is valid if and only if the eigenvalues of $C$ are within the disk of convergence of the usual scalar series for the logarithm... – J. M. May 13 '12 at 8:53
As for the Cauchy-like approach indicated by the second formula in Norbert's answer, see this. – J. M. May 13 '12 at 8:56

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.