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

I'm trying to calculate a Christoffel symbol, but I'm stuck on showing that $$\frac{\partial(e^{2A})}{\partial r}=2e^{2A}\frac{\partial A}{\partial r}\ ? $$

Nice easy steps would be much appreciated.

Thank you

share|cite|improve this question
What exactly is the problem? This is the chain rule. Assumming $A = A(r)$, $A$ is differentiable w.r. to $r$ and $e^x$ is the common exponential function. – user20266 Dec 5 '11 at 18:40
@Peter4075: Good! Then you've got it! Just take $u=2A$ in your case. You will have $du/dr=2 dA/dr$, and that's all there is to it. – Hans Lundmark Dec 5 '11 at 19:23
Dumb question: Is this a matrix exponential? Is that why you're worried? Best, – Dylan Moreland Dec 5 '11 at 20:32
Another dumb question then: how come you're calculating Christoffel symbols...? No offence, but that sounds quite advanced if you're having trouble with the chain rule. (P.S. For messages to another user, put "@" before the user name, and they will get notification in their inbox at the upper left of the page.) – Hans Lundmark Dec 5 '11 at 21:41
And by the way: perhaps if you write $u=A(r)$ it's easier to see why $r$ is the variable to differentiate with respect to; $A(r)$ just symbolizes any function of $r$, like $A(r)=r^5$ in the example. – Hans Lundmark Dec 5 '11 at 21:43

When $A$ is a scalar function of $r$, the formula $\frac{d }{d r}e^{2A}=2e^{2A}\frac{dA}{dr}$ is nothing but the chain rule.

When $A$ is a matrix, one has to watch out for the lack of commutativity. If $A$ and $\frac{dA}{dr}$ commute, everything works as for scalars. Otherwise, the formula breaks down: for example, if $$A(r)=\begin{pmatrix}0 & 1 \\ r & 0\end{pmatrix} \tag1$$ then $$\frac{d}{dr}_{\, \big|r=0} \exp A(r) = \begin{pmatrix}1/2 & 1/6 \\ 1 & 1/2\end{pmatrix}\tag2$$ which is not a product of $A'(0)=\begin{pmatrix}0 & 0 \\ 1 & 0\end{pmatrix}$ and $\exp A(0)=\begin{pmatrix}1 & 1 \\ 0 & 1\end{pmatrix}$, in either order: $$A'(0)\exp A(0) =\begin{pmatrix}0 & 0 \\ 1 & 1\end{pmatrix}\qquad (\exp A(0)) \,A'(0)=\begin{pmatrix}1 & 0 \\ 1 & 0\end{pmatrix}\tag3$$ One may observe in passing that the average of two products in (3) is $$\frac12 (A'(0)\exp A(0) + (\exp A(0)) \,A'(0)) = \begin{pmatrix}1/2 & 0 \\ 1 & 1/2\end{pmatrix} \tag4$$ which does a decent job of approximating (2). This suggests that (4) may be the beginning of a power series that gives the correct value of the derivative of matrix exponential, which (I guess) can be fished out of the Zassenhaus formula.

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.