6
$\begingroup$

Let $a, b \in (1, \infty)$ and $ m,n $ natural numbers at least equal to $2$ with $a\leq b$ and $m\leq n$.

Which is the largest of the numbers $$ A =(a^{\frac{1}{n}}+b^{\frac{1}{n}})^{\frac{1}{m}}$$ and $$ B=(a^{\frac{1}{m}}+b^{\frac{1}{m}})^{\frac{1}{n}} ?$$

We applied the inequality of generalized averages, calculations with radicals that did not lead to the expected response.

$\endgroup$
  • 2
    $\begingroup$ Well, what result do you expect? $\endgroup$ – Shaun Sep 28 '17 at 12:54
  • 3
    $\begingroup$ Dividing both of the expressions by $a^{1/mn}$, you can eliminate one variable. Might make the analysis easier. Also, raising both expressions to the power of $mn$ means all the $m$ appear in one expression, and all the $n$ appear in the other. Should make the analysis easier. $\endgroup$ – Arthur Sep 28 '17 at 12:59
  • $\begingroup$ Check Power Means Inequality... $\endgroup$ – Macavity Sep 28 '17 at 15:41
  • $\begingroup$ I don't want to be pendantic but you mean "larger"; not "largest". I only bring that up because I spent the first five minutes assuming you meant which was the largest possible values for $A$ and $B$ and they are unbounded as $a,b$ can be as large as you like. $\endgroup$ – fleablood Sep 29 '17 at 16:01
4
$\begingroup$

Note that $A\geq B$ iff $$a^{1/m}(1+x)^r=(a^{1/n}+b^{1/n})^{n/m}=A^{n}\geq B^{n}=(a^{1/m}+b^{1/m})=a^{1/m}(1+x^r)$$ where $x=(b/a)^{1/n}\geq 1$, and $r=n/m\geq 1$.

So it remains to show that $(1+x)^r\geq (1+x^r)$, that is $$\left(\frac{1}{1+x}\right)^r + \left( \frac{x}{1+x}\right)^r\leq 1$$ which holds because $t^r\leq t$ for $t\in[0,1]$.

$\endgroup$
  • $\begingroup$ @BarryCipra Thanks. Is it better now? $\endgroup$ – Robert Z Sep 28 '17 at 13:34
  • $\begingroup$ Thanks @Robert Z! With other calculations we have reached an inequality equivalent to $(1+x)^r \geq 1+x^r$ that I wanted to demonstrate without derivatives and that I did not succeed. Is it possible to prove no derivative? $\endgroup$ – medicu Sep 29 '17 at 15:31
  • $\begingroup$ @medicu See my edited answer. Any further doubt? $\endgroup$ – Robert Z Sep 29 '17 at 15:47

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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