Proof that winding number is integer - what does lifting mean?

728 Views Asked by At

Problem: Let $\alpha:[0,1] \rightarrow \mathbb{C}^*$ be a closed curve $\alpha(0)=\alpha(1)=1$. Then exists a unique $\tilde{\alpha}: [0,1] \rightarrow \mathbb{C}$ such that $\tilde{\alpha}(0)=0$ and $\exp \circ \tilde{\alpha} = \alpha$.

I am having trouble understanding the proof(italics part) - what exactly is meant?

Proof: For any point $a \in \mathbb{C}^*$ exists an open neighborhood $V$, such that $\exp^{-1}(V) = \bigsqcup_n U_n$, where each $U_n$ is homeomoprhically mapped onto $V$.

By Heine Borel, we deduce exictence of division $[0,1]$ such that $\alpha([a_n,a_{n+1}])$ is contained in an open set $V$ as above.

(NOT understand) We can then lift the whole curve, by inductively lifting pieces in the corresponding parallel horizontal strips such that they glue at initial and end points.

1

There are 1 best solutions below

0
On BEST ANSWER

So we have the subdivision of $[0,1]$ such that $\alpha([a_n,a_{n+1}])$ is contained in an open $V$ and we have a lift $\tilde{\alpha}_n:[a_n,a_{n+1}] \rightarrow \mathbb{C} $ such that the composition $\exp \circ \tilde{\alpha}_n([a_n,a_{n+1}]) = \alpha([a_n,a_{n+1}])$. Let us start with the partition $[0,a_1]$, then we have an open set, call it $V_0$, where $\alpha([0,a_1])$ has a lift and a collection $\bigsqcup_n U_n^0$ with each $U_n^0 \approx V_0$. Select one of these $U_n^0$ and call it $O_0$. Now, for the partition $[a_1,a_2]$, we have the open set $\alpha([a_1,a_2]) \subset V_1$ and collection $\bigsqcup_n U_n^1$. Notice each $U_n^1$ contains one of $\exp^-1(\alpha(a_1))$ and we choose $O_1=U_n^1$ such that $U_n^1 \cap O_0 \neq \emptyset$. Thus $\alpha([0,a_2])$ has a lift $\tilde{\alpha}:[0,a_2]\rightarrow \mathbb{C}$ defined by $$ \tilde{\alpha}(x) = \left\{ \begin{array}{cc}\tilde{\alpha}_0(x) & x \in [0,a_1] \\ \tilde{\alpha}_1(x) & x \in [a_1,a_2]\end{array}\right.$$

Repeat this process by induction to cover $[0,1]$ and we arrive at the lift $\tilde{\alpha}$ of $\alpha$. At the risk of repeating ad nauseam, that is $\tilde{\alpha}$ is such that $\exp \circ \tilde{\alpha} = \alpha$.