Help with excersice regarding inverse function theorem

68 Views Asked by At

I need help in solving the following exercise.

Consider a $\mathcal C^1$ function $f : U \subseteq \mathbb R^n \to \mathbb R$ with $U$ an open subset.

In the next exercise, it will be proved that $f$ cannot be injective. Let $\mathbf p \in U$ be a point for which $\nabla f(\mathbf p) \neq \mathbf 0.$ (If this point does not exist, then there is nothing to prove.) (Why?)

Let's say that $f_{x_1}(\mathbf p)\neq 0.$ Consider the function $F:U\subseteq \mathbb R^n \to \mathbb R^n$ defined by $F(\mathbf x)=F(x_1,x_2.....,x_n)=(f(\mathbf x),x_2, \dots, x_n).$ Show that this function has an inverse function $F^{-1}$ defined in some open ball of $F(\mathbf p) \in \mathbb R^n.$ Conclude from here that there exist infinitely many points $\mathbf x \in \mathbb R^n$ such that $f(\mathbf x)=c$ for some $c \in f \ \operatorname{rank}.$ This is to say, conclude that $f(\mathbf x)$ is not injective.

I was successful in proving that there exists some $F^{-1}$ defined in some open ball of $F(\mathbf p) \in \mathbb R^n.$ Therefore, there does exist $x_i = \phi_i(f(\mathbf x),x_2,x_3....x_n)$ for all $i.$ But I don't know how to continue from here to prove that $f(\mathbf x)$ is not injective. Please help. Thanks in advance.

1

There are 1 best solutions below

2
On

For the WHY point notice that if the gradient vanishes in all points the function has to be constant. For the rest notice that in a neighborhood of that point you can construct a hypersurface, all the points are mapped to the same value if the gradient doesn’t vanish.