3
$\begingroup$

Consider a continuous function $\alpha:[0,\infty) \to [0,\infty)$ satisfying the following properties:

  1. $\alpha(0)=0;$
  2. $\alpha$ is a strictly increasing function;
  3. $\alpha$ is smooth in $(0,\infty);$ and
  4. $|\alpha(x)/x| \to \infty$ as $x\to 0^+$.

Under these conditions, is it true that there exists $\delta>0$ and $Μ >0$ such that $$ \left|\frac{\alpha(x)}{\alpha(x/2)}\right| \leq M,\ \forall\ x\in(0,\delta)\ ?$$

I do not know if the above question is either true or false. Could someone please help me to solve this question?


Searching for a counter-example:

Since I could not prove the above question, I started looking for a counter-example. However, the examples that I was able to find that satisfy 1-4, such as

  1. $\alpha(x) = x^\beta$, for $0<\beta<1$;
  2. $\alpha(x) = -x \log (x)$ in a neighborhood of $0;$ and
  3. $\alpha(x) = (1-x^x)/x$ in a neighborhood of $0.$

However, all these functions fulfil the desired property.

$\endgroup$
2
  • $\begingroup$ Maybe you want an additional requirement on the behavior near zero. Else you can use a piecewise defined function, like $\alpha(0)=0$ and $\alpha(x)=x+1$ for $x>0$. $\endgroup$ Commented Feb 7, 2023 at 0:13
  • $\begingroup$ Yeah, this is correct. Thx $\endgroup$ Commented Feb 7, 2023 at 9:56

1 Answer 1

5
+200
$\begingroup$

No. Let $\phi:\mathbb R \to [0, 1]$ be a smooth increasing step function, satisfying $\phi(x) = 0$ for $x \le 1$, and $\phi(x) = 1$ for $x \ge 2$. (For example, take the integral of a bump function.)

Then let $a_n = n^{-n}$, and define

$$\alpha(x) = x^{1/2} + \sum_{n=1}^\infty \left(a_n^{1/2} - a_{n+1}^{1/2}\right) \phi\left(\frac{x}{a_{n+1}}\right).$$

You can check that this function is well-defined, and satisfies your conditions, but we have

$$\frac{\alpha(2a_{n+1})}{\alpha(a_{n+1})} \ge \frac{a_n^{1/2}}{2a_{n+1}^{1/2}} \ge \frac{n^{1/2}}2.$$

(As an aside, the result is true if $\alpha(0) = 0$ and $\alpha$ is concave and increasing, which seems like the intuition the conditions are aiming for. This counterexample works because it's not a concave function.)

$\endgroup$

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.