Here's the question you clicked on:
Ted.knows.it
Prove the following inequality: equation attached.
\[\ln \frac{ x_{1} + x_{2} + x_{3} + ... + x_{n} - n }{ x_{1} + x_{2} + x_{3} + ... + x_{n} } \ge \frac{ \ln ((x_{1}-1)(x_{2}-1)...(x_{n}-1)) - \ln(x_{1} x_{2} .... x_{2}) }{ n }\]
It is the point c) of a math analysis problem which gives the function: \[f : (1, \infty) \rightarrow \mathbb{R} \] \[f(x) = \ln \frac{ x-1 }{ x }\]
I managed to bring it to the following form: \[(\frac{ x_{1}-1+x_{2}-1+...+x_{n}-1 }{ x_{1}+x_{2}+...+x_{n} } )^{n} \ge \frac{ (x_{1}-1)(x_{2}-1)...(x_{n}-1) }{ x_{1} x_{2} ... x_{n}}\]
In your proff you are obliged to use this function?\[f(x) = \ln \frac{ x-1 }{ x }\]
sorry for the late answer.
no, they say nothing about using it.
\[\ln \frac{ x_{1} + x_{2} + x_{3} + ... + x_{n} - n }{ x_{1} + x_{2} + x_{3} + ... + x_{n} }=\ln\frac{\frac{\sum_{i=1}^n(x_i-1)}{n}}{\frac{\sum_{i=1}^n x_i}{n}}=\ln\frac{\sum_{i=1}^n(x_i-1)}{n}-\ln\frac{\sum_{i=1}^n x_i}{n}\]Analogically:\[\frac{ \ln ((x_{1}-1)(x_{2}-1)...(x_{n}-1)) - \ln(x_{1} x_{2} .... x_{2}) }{ n }=\ln\sqrt[n]{\prod_{i=1}^n(x_i-1)}-\ln\sqrt[n]{\prod_{i=1}^n x_i}\]Now use somehow that the arithmetic mean is bigger than the geometric mean:\[\frac{\sum_{i=1}^n x_i}{n}\geqslant \sqrt[n]{\prod_{i=1}^n x_i} .\]
Hm.. This is not very useful.
i tried to analyse the problem step by step and started with x1 \[\ln \frac{ x_{1} - 1 }{ x_{1} } \ge \frac{ \ln (x_{1} - 1) - \ln x_{1} }{ 1 }\] which is true.
however, i think that the mathematical induction is not the most viable solution in this case.
Ok. What is next? \[\ln\frac{x_1+x_2-2}{x_1+x_2}\geqslant\frac{\ln\left((x_1-1)(x_2-1)\right)-\ln x_1x_2}{2}\]Are you able to prove it?
one question...i want make myself sure :) \[x_i>1\]?
I have a way to prove this using the inequality for arithmetic and geometric means.
now i was looking at that idea. it seems the right one...
@mukushla , yes it is given by the function domain
And also you need to know that \(f(x)=\ln\frac{x-1}{x}\) is an increasing function. Think about the idea I wrote above. It is not so difficult.
yeah, so all in all, the problem goes down to proving that arithmetic mean is bigger than the geometric mean..
Yes. I wish you luck while proving this one.
thanks for your time. :)
this one from me...i hope it will help\[f''(x)=\frac{-2x+1}{x^2(x-1)^2}<0 \ \ , \ \ \forall x>1\]so \(f\) is concave. Then using the Jensen’s inequality we obtain:\[\ln(1-\frac{1}{\frac{x_1+x_2+...+x_n}{n}}) \ge \frac{1}{n} \sum_{i=1}^{n} \ln(1-\frac{1}{x_i})\]we are done