1
$\begingroup$

I had a question about differential operators. I give some context. The operator \begin{align} x\partial_x \end{align} has as differential symbol the function $x\xi$ (replacing $\partial_x$ by $\xi$)

Another way to obtain the symbol $P(x,\xi)$ of an arbitrary differential operator $P(x,\partial_x)$ is by

\begin{align} P(x,\xi)=e^{-x\xi}P(x,\partial_x)e^{x\xi}(1) \end{align} where 1 is the function identically 1.

Given the above, with $P(x,\partial_x)=x\partial_x$, as

\begin{align} P(x,\xi)&=e^{-x\xi}x\partial_x e^{x\xi}(1)\\ &=e^{-x\xi}x\xi e^{x\xi}(1)\\ &=e^{-x\xi+x\xi}x\xi(1)\\ &=e^{0}x\xi(1)\\ &=x\xi \end{align}

Question 1. Does the operator $e^{\partial_x}$ have the symbol $e^{\xi}$? My attempt: \begin{align} e^{-x\xi}e^{\partial_x}e^{x\xi}(1)&=e^{-x\xi}\sum_{k=0}^{\infty}\frac{\partial_x^k}{k!}e^{x\xi}(1)\\ &=e^{-x\xi}\sum_{k=0}^{\infty}\frac{\xi^k}{k!}e^{x\xi}(1)\\ &=e^{-x\xi}e^{\xi}e^{x\xi}(1)\\ &=e^{\xi} \end{align}

Another way to calculate the symbol is with the Baker-Campbell formula (https://en.wikipedia.org/wiki/Baker–Campbell–Hausdorff_formula

My attempt:

\begin{align} e^{-x\xi}e^{\partial_x}e^{x\xi}(1)&=e^{-x\xi}e^{\partial_x+x\xi+\frac{1}{2}\xi}(1)\\ &=e^{(-x\xi+\partial_x+x\xi+\frac{1}{2}\xi)+\frac{1}{2}\xi}(1)\\ &=e^{\partial_x+\xi}(1)\\ &=\left.\sum_{k=0}^{\infty}\frac{(\xi+\partial_x)^k}{k!}(f)\right|_{f=1}\\ &=e^{\xi} \end{align}

Question 2. What should be the Lie group and Lie algebra to apply this formula? I don't know what they are. I guess that

\begin{align} \text{Lie algebra}=\left\{aI,b\partial_x:a,b\in\mathbb{R}\right\}\\ \text{Lie Group}=\left\{e^{aI}, e^{b\partial_x}:a,b\in\mathbb{R}\right\}? \end{align} Thanks.

$\endgroup$
5
  • $\begingroup$ Q1 is correct. Q2: By the BCH formula $\{e^{b\,\partial_x}:b\in\mathbb R\}\cong\mathbb R$ is a group and obviously also a (quite boring) Lie group. Its equally boring Lie algebra should now be obvious. $\endgroup$
    – Kurt G.
    Commented Aug 30, 2023 at 6:17
  • $\begingroup$ Thank you so much. What I have doubt is the following. First you verify that a group is a Lie group and then you find the Lie algebra? I ask it because in this case, it seems more intuitive to me to find the algebra of lie (formed by the fields) and then with the exponential form the group of lie. Is it possible to first define a lie algebra and then find the associated lie group? thank you. $\endgroup$
    – eraldcoil
    Commented Aug 30, 2023 at 6:21
  • $\begingroup$ In this case yes. They key to all of this is the BCH formula in connection with $[\partial_x,\partial_x]=0\,.$ So you can start from this trivial Lie algebra which is isomorphic to $\mathbb R\,.$ You may want to to study more interesting examples. $\endgroup$
    – Kurt G.
    Commented Aug 30, 2023 at 6:29
  • $\begingroup$ I will try to understand what you say. From what I understood, the existence of the exponential application is the one that given a Lie Algebra $\mathcal{G}$ then $\exp(\mathcal{G})$ will be a Lie group? and furthermore, this Lie group have $\mathcal{G}$ as the Lie algebra? This is the idea? $\endgroup$
    – eraldcoil
    Commented Aug 30, 2023 at 7:55
  • 1
    $\begingroup$ Please do not ask me to explain the bare basics about Lie algebras and Lie groups in this comment section. There are plenty of resources which do that better. To add some juicy meat I recommend to study standard Lie groups such as $SU(2),SO(3)$ and their algebras. $\endgroup$
    – Kurt G.
    Commented Aug 30, 2023 at 12:00

0

You must log in to answer this question.

Browse other questions tagged .