0
$\begingroup$

Suppose we have $$e^{At}e^{Bt}=F(t),$$ where $$A, B$$ - operators that do not commute. Now I need to take the derivative $$dF(t)/dt.$$ In which order do I write the operators?

$$dF(t)/dt = Ae^{At}e^{Bt} + e^{At}Be^{Bt}$$ or $$dF(t)/dt = e^{At}Ae^{Bt} + e^{At}e^{Bt}B~?$$

$\endgroup$
0

2 Answers 2

5
$\begingroup$

They are the same, since any operator commutes with its exponential \begin{align} Ae^{tA} &= A\left(1+tA+\frac{1}{2!}t^2A^2+\dots\right) \\ &= A+tA^2+\frac{1}{2!}t^2A^3+\dots \\ &= \left(1+tA+\frac{1}{2!}t^2A^2+\dots\right)A \\ &= e^{tA}A \end{align} (and, in general, any operator $A$ commutes with every function $f(A)$ of it, for similar reasons).

$\endgroup$
1
  • $\begingroup$ Thank you very much, your explanation is very helpful! $\endgroup$
    – prividenie
    Commented Jun 8, 2019 at 19:35
1
$\begingroup$

Expanding on Cosmas Zachos' comment Suppose $G(x)$ is some function which has an expression as a Taylor series.

$G(x) = \sum_i g_i x^i$

It can be proven* by induction that

$$ [X,G(X)]=0 $$

This means $[X,e^{Xt}$]=0$

In particular

$$ Ae^{At} = e^{At}A $$ $$ Be^{Bt} = e^{Bt}B $$

so both of your expressions are equal. What is important is to take care that you don't swap any terms involving $A$'s with any terms involving $B$'s. The answer may have been more complicated.

*At least in a physics usage of the word..

$\endgroup$
1
  • $\begingroup$ Thank you very much for your answer! $\endgroup$
    – prividenie
    Commented Jun 8, 2019 at 19:36

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.