MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Assume that $A$ is an arbitrary positive integrable function on $[0,1]$. Whether exists a convex function $f_A(x)=x g(x)$ of $(0,+\infty)$ into itself (depending on $A$) such that $\lim_{x\to +\infty} g(x)=+\infty $ and $$\int_0^1 A(x) g(1/x^2) dx <+\infty.$$ This question is related to membership of $g$ to some Dini class.

share|cite|improve this question
up vote 2 down vote accepted

The answer is yes. First define inductively a sequence $x_k>0$ such that $x_{k+1} < x_k/2$, and $$\int_0^{x_k}A(x)dx<2^{-k}.$$ This is possible because $A$ is integrable.

Then define a continuous function $[0,1]$ by $h(x_k)=k$ and $h$ is linear on each interval $[x_{k+1},x_k]$. It is easy to see that this function is convex, decreasing and tends to $+\infty$ as $x\to 0+$. Moreover $$\int_0^1A(x)h(x)dx\leq\sum (k+1)2^{-k}<\infty.$$ Now set $g(x)=h(1/\sqrt{x})$ and it remains to verity that $xg(x)$ is convex. This we do by differentiation: $$(xg(x))^\prime=h(x^{-1/2})-2x^{-1/2}h'(x^{-1/2}).$$ Both summands are increasing, therefore $xg(x)$ is convex.

share|cite|improve this answer
Very nice solution. – djoke Nov 30 '12 at 13:21

Let's change variables: we have: $$\int_0^1A(x)dx=\frac{1}{2}\int_1^{+\infty} A(x^{-1/2}) x^{-3/2}dx < +\infty\, .$$

Therefore, by the Dominated Convergence Theorem $$\int_1^{+\infty} A(x^{-1/2}) x^{-5/2}(x-k) _ +dx=o(1),\qquad (\mathrm{as }\, k\to+\infty)\, ,$$ and in particular there is a sequence $k _ n\to +\infty$ such that $$\int_1^{+\infty} A(x^{-1/2}) x^{-5/2}(x-k _ n) _ +dx\le 2^{-n} .$$ The function $f:(0,+\infty)\rightarrow(0,+\infty)$ $$f(x):=x+\sum_{n=1}^\infty(x-k_n) _ +$$ is convex and verifies $g(x):=f(x)/x\to+\infty$; moreover, integrating by series $$\int_1^{+\infty} A(x^{-1/2}) g(x) x^{-3/2} dx < +\infty\, ,$$ and by the same change of variable as before, the latter integral is twice $$\int_0^1A(x) g(1/x^2) dx\, .$$

share|cite|improve this answer
Thanks for the solution. – djoke Nov 30 '12 at 13:21

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.