I'm trying to a find an onto function $f:\mathbb{R}_{\geq 0} \to \mathbb{R}_{\geq 0}$ such that the $f^{(k)}(0)=0$ for all $k \geq 0$ I'm familiar with functions such as $e^{-1/x^2}$ but the main problem I'm having is finding one that is onto. I suspect it needs to be some kind of infinite series of functions (to ensure the derivative is always zero at zero for all derivatives), but I'm truly lost on how to proceed. Most attempts I make eventually end up with a $k$th derivative that's non-zero at zero (such as $e^x-1$) or isn't onto (such as $e^{-1/x^2}$) and I really have no idea how to solve this issue. Any help or hints would be sincerely appreciated. Thanks.
Asked
Active
Viewed 60 times
1
-
1$\tan (\frac {\pi} 2 e^{-1/x^{2}})$ does the job. – geetha290krm Sep 16 '22 at 05:01
-
@geetha290krm: isn't this function not defined at $x = 0$? – user1090793 Sep 16 '22 at 05:11
-
@user1090793 As in the case of $e^{-1/x^{2}}$ you just define it to be $0$ when $x=0$. – geetha290krm Sep 16 '22 at 05:14
2 Answers
2
Consider
$$f(x) =\frac{e^{-1/x}}{1-e^{-1/x}}$$ for $x>0$, and $0$ for $x\le 0$.
The idea is that the function $e^{-1/x}$ maps $[0, \infty)$ bijectively onto $[0,1)$, and $\frac{t}{1-t}$ maps $[0, 1)$ onto $[0, \infty)$.

orangeskid
- 53,909
-
1
-
@geetha290krm: Yes, indeed, I noticed that. Our functions are both convex. The third derivative changes sign. Now, we know that it's not possible to have all derivatives $\ge 0$, since then the function has to be analytic on $\mathbb{R}$ . – orangeskid Sep 16 '22 at 06:57
1
Here's an idea based on smooth transition functions. Take $g(x)$ to be the function defined in the accepted answer to this question:
How to build a smooth "transition function" explicitly?
Now define a piecewise function $t(x): \mathbb{R}_{\geq 0} \to \mathbb{R}_{\geq 0}$ by:
$$t(x) = g(x - n) + n \text{ on the interval } [n, n + 1), n \in \mathbb{N}$$
Since $g(x)$ is smooth, $g^{(0)}(0) = 0$, and $g^{(k)}(0) = g^{(k)}(1) = 0$ for $k > 0$, $t(x)$ is smooth and $t^{(k)}(0) = 0$ for all $k \geq 0$. Clearly, $t(x)$ is surjective.

user1090793
- 1,053