# Prove that $a^{\frac{2n}{m}} - 1 \geq n\big(a^{\frac{n+1}m} - a^{\frac{n-1}{m}}\big)$.

Given natural numbers $m,n,$ and a real number $a>1$, prove the inequality :

$$\displaystyle a^{\frac{2n}{m}} - 1 \geq n\big(a^{\frac{n+1}m} - a^{\frac{n-1}{m}}\big)$$

SOURCE : Inequalities (PDF) (Page Number 2 ; Question Number 153.2)

I have been trying this problem from 2 weeks but still no success. I tried every method I could think of like AM-GM, C-S, Holder and more, but could not find a proof.

Also, is it necessary for $n,m$ to be natural numbers ?

Any help will be gratefully acknowledged.

The answer can be given via simple calculus, and thus the result can be shown to hold true for all $x$. However, the OP has stated that he/she would prefer a solution that did not resort to calculus. So here is my edited answer. For my original answer, please check the edit history.

Let $a^{\frac{1}{m}}=x>1$. The question is equivalent to showing $$x^{2n}-1 \ge n(x^{n+1}-x^{n-1}) \iff \frac{x^{2n}-1}{x^2-1} \ge nx^{n-1}$$ Now, note that $$\frac{x^{2n}-1}{x^2-1}=\sum_{k=0}^{n-1}x^{2k}=\frac{1}{2} \left(\sum_{k=0}^{n-1}x^{2k}+x^{2n-2k-2}\right) \ge \frac{1}{2} \times 2\sum_{k=0}^{n-1}x^{n-1}=nx^{n-1}$$ From $\text{AM-GM}$. Our proof is done.

• Isn't $g(x) = x^{2n} - nx^{n+1}+ nx^{n-1}-1$ ? That gives a different derivative. Am I overlooking something? – Martin R Feb 16 '17 at 9:36
• @MartinR Right, I missed a step. Does it work now? – S.C.B. Feb 16 '17 at 9:39
• @S.C.B. Any idea of a proof without calculus.... – user399078 Feb 16 '17 at 9:40
• @Nirbhay Do we have to do a proof without calculus? – S.C.B. Feb 16 '17 at 9:40
• I still cannot verify your expression for $g'(x) = f''(x)$ ... – Martin R Feb 16 '17 at 9:42

Let $x = a^{\frac 1m} > 1$. Using $$x^{2n} - 1 = (x-1)(1+ x+x^2 + \ldots + x^{2n-1}) \\ x^{n+1} - x^{n-1} = (x-1) (x^{n-1}+x^n)$$ we get $$x^{2n} - 1 - n(x^{n+1} - x^{n-1}) = (x-1)\left( 1+ x+x^2 + \ldots + x^{2n-1} - n(x^{n-1}+x^n) \right) \\ = (x-1)\sum_{k=1}^n \left( x^{k-1} + x^{2n-k} - x^{n-1}-x^n\right) \\ = (x-1)\sum_{k=1}^n (x^{n-k}-1)(x^n - x^{k-1}) \\ \ge 0$$ (with strict inequality for $n \ge 2$), which is the desired inequality $$x^{2n} - 1 \ge n(x^{n+1} - x^{n-1}) \, .$$ This proof works for positive real $m$ and integer $n \ge 1$. For $x < 1$ the same inequality with $\ge$ replaced by $\le$ holds.

• Shoot, I first downvoted, and then I removed it, but there must have been an error. Sorry. – S.C.B. Feb 16 '17 at 12:28