0

I am working towards, seeing how the ‘Lie Derivative’ is used to look at symmetry, in General Relativity.

I do find, that I currently have difficulty with reading the pure maths involved.

I found an exercise, at

Why do we need a Lie derivative of a vector field?

that seems at the level that might be useful to me ( see after paragraph ‘3’). Currently though (05th Sept. 2022) the above post does not contain the solution to the exercise. Please note my various comments on that post.

I would like a reference to some simple ‘Exercises with Solutions’, on finding Lie derivatives.

Exercises, which do not involve the use of the commutator definition of the Lie derivative, $L_{Y} X$, as the main idea, i.e.

\begin{equation*} L_{Y} X=\left[ Y, X \right] \end{equation*} where $Y$ and $X$ are vector fields.

Please note the Lie derivative can be extended to other mathematical objects, it is not only used for vector fields.

1 Answers1

2

Well, here's a more detailed account of how to solve that exercise:

You have $Y=-y\dfrac{\partial}{\partial x} +x\dfrac{\partial}{\partial y}$

To compute its flow $\phi^Y_t$ you need to solve the following set of coupled ODE's:

$$\dot{x}=-y$$ $$\dot{y}=x$$

This is not too hard to solve (just differentiate one of the equations and plug in the other one, you'll get a Harmonic oscillator-esque ODE). In any case the solution is (which you can explicitly check by plugging in):

$$x(t)=x_0cos(t)-y_0sin(t)$$

$$y(t)=x_0sin(t)+y_0cos(t)$$

Hence the flow is $$\phi^Y_t(x,y)=\big(xcos(t)-ysin(t),xsin(t)+ycos(t)\big)$$

I switched from $x_0,y_0$ to $x,y$ since the flow can take arbitrary initial conditions $x,y$ and trace out their evolution as a function of $t$.

Now compute $d\phi^Y_t$ in standard coordinates (i.e. in the identity chart); its matrix representation will be just the Jacobian matrix of partial derivatives hence:

$$ d\phi^Y_t= \begin{pmatrix} cos(t) & -sin(t) \\ sin(t) & cos(t) \\ \end{pmatrix} $$

Hence $$(d\phi^Y_t)_{\phi^Y_{-t}(p)} \bigg(\epsilon \dfrac{\partial}{\partial x}\bigg)_{\phi^Y_{-t}(p)}=\begin{pmatrix} cos(t) & -sin(t) \\ sin(t) & cos(t) \\ \end{pmatrix}\begin{pmatrix} \epsilon \\ 0 \\ \end{pmatrix}=\epsilon\begin{pmatrix} cos(t) \\ sin(t) \\ \end{pmatrix}$$

The subscripts indicate "evaluated at" hence $(d\phi^Y_t)_{\phi^Y_{-t}(p)}$ means that it is $d\phi^Y_t$ evaluated at the point $\phi^Y_{-t}(p)$ (notice that it is constant as a function of $x,y$ hence we shouldn't worry about that but I included it for completeness).

$$(L_Y X)_p=\dfrac{d}{dt}\bigg|_{t=0} \bigg((d\phi^Y_t)_{\phi^Y_{-t}(p)} \bigg(\epsilon \dfrac{\partial}{\partial x}\bigg)_{\phi^Y_{-t}(p)}\bigg)=\begin{pmatrix} 0 \\ \epsilon \\ \end{pmatrix}$$

The reason I included $\epsilon$ should be clear from the post in your link; Also you can use other charts to compute $d\phi^Y_t$; for instance in the polar chart it would be the identity.

Leonid
  • 1,679