Properties of a surjective local diffeomorphism

Question

Assume that $f:\mathbb{R}^N\to\mathbb{R}^N$ is a surjective function and in addition suppose that $f$ is a local diffeomorphism. Take two points in the image of $f$, let's say, $f(x),f(y)$ with $f(x)\neq f(y)$.

Is it possible to find a continuous curve $\alpha :[0,1]\to\mathbb{R}^N$ such that $\alpha(0)=x$, $\alpha(1)=y$ and $f(\alpha (t))=(1-t)f(x)+tf(y)$.

Remark: This question is related to this one

Remark 1: The answer given here by @smiley06, solves this problem, when only one boundary condition is prescribed.

Thank you

You're basically asking whether $f$ has to be a covering. For $N\geq2$ it is not true. — user8268, Sep 25 '13 at 12:51
Thank you for your comment @user8268. Could you please give me a example? — Tomás, Sep 25 '13 at 12:52
Do you have any example of such a $f$ that is not invertible? I can't think of any. — tom, Sep 25 '13 at 12:54

score 3 · Answer 1 · answered Sep 25 '13 at 13:52

3

All the credit goes to the user8268 but I made a little animation.

enter image description here

answered Sep 25 '13 at 13:52

tom

4,596

Unfortunately I don't know how to fill the inside of the disk. – tom Sep 25 '13 at 13:52
Thank you @tom for your kindness. Now I can see, hahah. – Tomás Sep 25 '13 at 14:10
wow! How did you do it? – user8268 Sep 25 '13 at 14:11
I used Autodes Maya, but it is "a little" bit overkill(most of the Avatar scenes were done in it :D). I guess it can be done in something like Flash. – tom Sep 25 '13 at 14:25

score 2 · Accepted Answer · answered Sep 25 '13 at 13:07

2

For $N=2$; I replace $\mathbb R^2$ with a disk (as it is diffeomorphic); here is a surjective map from a disk to a "disk" (still diffeomorphic to $\mathbb R^2$) and a straight segment that can't be lifted.

really horrible picture

answered Sep 25 '13 at 13:07

user8268

21,348

Sorry about my dumbness, but could you add some explanation. Thank you – Tomás Sep 25 '13 at 13:11
I know that picture is impossible; try yourself to map a disk to the plane so that the image is a disk, but the map is not 1-1. – user8268 Sep 25 '13 at 13:17
Very nice! I like it. – tom Sep 25 '13 at 13:21
Are all that "loops" necessary? Disclaimer: I am not specialist. – Matemáticos Chibchas Sep 25 '13 at 13:41
Yes the loops are necessary. Look at my animation. You take disk first deform it to the O shape and with one end you fill the hole in the O. – tom Sep 25 '13 at 13:59

Properties of a surjective local diffeomorphism

2 Answers2

Linked