Difference quotient of global flow and the lie derivative

32 Views Asked by At

I am trying to follow a proof and I am not so sure with a step there.

So, let $X\in\Gamma(TM)$ be a vector field with compact support on a $\partial$-manifold $M$ and $\psi^X_t$, $t\in[-1,1]$ the flow generated by $X$.

Let $\Sigma^X_t$ be the approximate Lie derivative, i.e. the difference quotient $$\Sigma^X_t:L^2\Omega^k(M)\to L^2\Omega^k(M),\quad \Sigma^X_t\omega = \frac{1}{t}\big((\psi^X_t)^*\omega - \omega\big).$$

Then $$\lim\limits_{t\to 0}\Sigma^X_t\omega = \mathcal{L}_X\omega.$$

Now in the proof I am trying to follow, they say that we can write $$\Sigma^X_t\omega = \int_0^1\mathcal{L}_X\bigg((\psi^X_{ht})^*\omega\bigg)dh.$$

I don’t quite understand that. Is it just that the fundamental theorem of calculus is applied?

Does applying the chain rule yield the following? $$\frac{d}{dh}\bigg((\psi^X_{ht})^*\omega\bigg) = \frac{d}{dh}(ht) \cdot \mathcal{L}_X\bigg((\psi^X_{ht})^*\omega\bigg) = t \cdot \mathcal{L}_X\bigg((\psi^X_{ht})^*\omega\bigg) $$ Why the Lie derivative for the outer derivative?

And then from that it follows the “claim” with the fundamental theorem of calculus?

$$\int_0^1\mathcal{L}_X\bigg((\psi^X_{ht})^*\omega\bigg)dh = \frac{1}{t}\int_0^1 \frac{d}{dh}\bigg((\psi^X_{h\cdot t})^*\omega \bigg)dh = \frac{1}{t}\big((\psi^X_{1\cdot t})^*\omega -(\psi^X_{0\cdot t})^*\omega\big) = \frac{1}{t}\big((\psi^X_{t})^*\omega -\omega\big).$$

Thanks in advance!

——————————

EDIT: I guess I figured it out:

$$\mathcal{L}_X((\psi^X_h)^*\omega)= \lim\limits_{t\to 0}\Sigma^X_t\big((\psi^X_h)^*\omega\big) = \lim\limits_{t\to 0}\frac{1}{t}\big((\psi^X_t)^*\big((\psi^X_h)^*\omega\big) - (\psi^X_h)^*\omega\big) = \lim\limits_{t\to 0}\frac{1}{t}\big((\psi^X_{t+h})^*\omega- (\psi^X_h)^*\omega\big) = \frac{d}{dh} \big((\psi^X_h)^*\omega\big). $$

Is that correct?