2
$\begingroup$

For all $n\in\mathbb N$ prove that: $$\frac{(n+2)^{n+1}}{(n+1)^n}-\frac{(n+1)^n}{n^{n-1}}

We can rewrite it in the following form $$(n+2)\left(1+\frac{1}{n+1}\right)^{n}-(n+1)\left(1+\frac{1}{n}\right)^{n-1}

  • 0
    Maybee take logarithms and use derivatives and standard calculus methods.2017-01-24
  • 0
    @Ahmed S. Attaalla $e$ does not give.2017-01-24
  • 0
    @MichaelRozenberg: Have you seen my answer below? I think mine is the complete proof. It actually partly cites your proof of convexity.2017-01-26

2 Answers 2

2

Claude Leibovici's answer prove only that the proposition is true for sufficiently large rather than arbitrary positive $n$. Here is the complete proof with stronger result.


Let $$f(x):=x\left(1+\frac{1}{x}\right)^{x},\ x>0.$$ $$f'(x) = e^{x\ln(1+\frac1x)}\Big[1+x\Big(\ln\Big(1+\frac1x\Big)-\frac1{1+x}\Big)\Big].$$ $f$ increases since $$\ln\Big(1+\frac1x\Big)=-\ln\Big(1-\frac1{1+x}\Big)>\frac1{1+x}$$ by Taylor expansion of the logarithmic function around $1$. Because $f(x)$ is strictly convex, $1=f'(0)

So we conclude more than the required inequality $$1

  • 0
    Thank you very much! It quits that I was near this solution.2017-01-26
  • 0
    This is the proof, indeed ! Thanks for providing it.2017-01-27
  • 0
    @ClaudeLeibovici and Michaeal Rozenberg: No worries. I am glad you both like it.2017-01-27
4

I do not know if this is a satisfactory answer for you.

Consider $$A=\frac{(n+2)^{n+1}}{(n+1)^n}\qquad , \qquad B=\frac{(n+1)^n}{n^{n-1}}$$ Take logarithms and expand as Taylor series for large values of $n$. This would lead to $$\log(A)=1+\log \left(n\right)+\frac{1}{2 n}+\frac{1}{3 n^2}-\frac{13}{12 n^3}+O\left(\frac{1}{n^4}\right)$$ $$\log(B)=1+\log \left(n\right)-\frac{1}{2 n}+\frac{1}{3 n^2}-\frac{1}{4 n^3}+O\left(\frac{1}{n^4}\right)$$ Taylor again $$A=e^{\log(A)}=e n+\frac{e}{2}+\frac{11 e}{24 n}-\frac{43 e}{48 n^2}+O\left(\frac{1}{n^3}\right)$$ $$B=e^{\log(B)}=e n-\frac{e}{2}+\frac{11 e}{24 n}-\frac{7 e}{16 n^2}+O\left(\frac{1}{n^3}\right)$$ $$A-B=e-\frac{11 e}{24 n^2}+O\left(\frac{1}{n^3}\right)$$

  • 0
    Beautiful way! Learned something. Thank you!2017-01-24
  • 1
    This only concludes the inequality for sufficiently large but not for arbitrary $n$. More accurate calculation is needed to draw the desired conclusion.2017-01-24
  • 0
    @Hans. I totally agree with you. I was hoping that my answer did precise that this was done for large values of $n$.2017-01-24
  • 0
    Please take a look at my proof of stronger result for arbitrary $n$.2017-01-26