If \(x,y\) are real numbers such that \(x^{2014}+y^{2014}=1\) then prove that:

\[\left(\sum_{n=1}^{1007} \dfrac{x^{2n}+1}{x^{4n}+1} \right) \left(\sum_{n=1}^{1007} \dfrac{y^{2n}+1}{y^{4n}+1} \right)< \dfrac{1}{(1-x)(1-y)}\]

If \(x,y\) are real numbers such that \(x^{2014}+y^{2014}=1\) then prove that:

\[\left(\sum_{n=1}^{1007} \dfrac{x^{2n}+1}{x^{4n}+1} \right) \left(\sum_{n=1}^{1007} \dfrac{y^{2n}+1}{y^{4n}+1} \right)< \dfrac{1}{(1-x)(1-y)}\]

No vote yet

1 vote

×

Problem Loading...

Note Loading...

Set Loading...

## Comments

Sort by:

TopNewestThere is a typo in the question. I will explain you what the typo is.

For suppose, take \(x=-1\) and \(y=0\). Then we get

\[ \sum_{n=1}^{1007} \dfrac{x^{2n}+1}{x^{4n}+1} = 1007 \\ \sum_{n=1}^{1007} \dfrac{y^{2n}+1}{y^{4n}+1} = 1007 \\ \dfrac{1}{\left(1-x \right) \left(1-y \right)} = \dfrac{1}{2} \]

So, according to our question it forces that \(1007 \times 1007 < 1/2\), which is impossible.

The question should be actually like this:

If \(x\), \(y\)are real numbers such that \(x^{2014}+y^{2014}=1\) then prove that:

\[\left( \sum_{n=1}^{1007} \dfrac{x^{2n}+1}{x^{4n}+1} \right) \left( \sum_{n=1}^{1007} \dfrac{y^{2n}+1}{y^{4n}+1} \right) < \dfrac{1}{\left(1-|x| \right) \left( 1-|y| \right)}\]

PROOF:Let \(|x| = a\) and \(|y| = b\). So, the given constraint changes into \(a^{2014} + b^{2014} = 1\).

Since, \( a^{4n} + 1 > \dfrac{ \left( a^{2n} + 1 \right)^2}{2} = \dfrac{1}{2} \times \left( a^{2n} + 1 \right) \times \left( a^{2n} + 1 \right) > a^{n} \left( a^{2n} + 1 \right)\). Which implies that

\[ \dfrac{1+a^{2n}}{1+a^{4n}} < \dfrac{1}{a^n} \\ \sum_{n=1}^{1007} \dfrac{1+a^{2n}}{1+a^{4n}} <\sum_{n=1}^{1007} \dfrac{1}{a^n} = \left( \dfrac{1}{a^{1007}} - 1 \right) \times \left( \dfrac{1}{1-a} \right)\]

Similary \(\sum_{n=1}^{1007} \dfrac{1+b^{2n}}{1+b^{4n}} < \left( \dfrac{1}{b^{1007}} - 1 \right) \times \left( \dfrac{1}{1-b} \right)\)

Multiplying the two, we get

\[ \sum_{n=1}^{1007} \dfrac{1+a^{2n}}{1+a^{4n}} \times \sum_{n=1}^{1007} \dfrac{1+b^{2n}}{1+b^{4n}} < \left( \dfrac{1}{a^{1007}} - 1 \right) \times \left( \dfrac{1}{b^{1007}} - 1 \right) \times \dfrac{1}{\left(1-a \right) \left( 1-b \right)}\]

So, If we prove that \( \left( \dfrac{1}{a^{1007}} - 1 \right) \times \left( \dfrac{1}{b^{1007}} - 1 \right) < 1\), then we are done.

Since \(a^{2014} + b^{2014} = 1 \) and \(0 \leq a, b \). It implies \(a,b < 1\). So, \(a^{1007} > a^{2014}\) and \(b^{1007} > b^{2014}\). Which implies that \(a^{1007} + b^{1007} > a^{2014} + b^{2014} = 1\).

So,

\[1-a^{1007} - b^{1007} <0 \\ \left( 1- a^{1007} \right) \times \left(1- b^{1007} \right) < a^{1007} \times b^{1007} \\ \left( \dfrac{1}{a^{1007}} - 1 \right) \times \left( \dfrac{1}{b^{1007}} - 1 \right) < 1 \]

Therefore,

\[\left( \sum_{n=1}^{1007} \dfrac{x^{2n}+1}{x^{4n}+1} \right) \left( \sum_{n=1}^{1007} \dfrac{y^{2n}+1}{y^{4n}+1} \right) = \left( \sum_{n=1}^{1007} \dfrac{a^{2n}+1}{a^{4n}+1} \right) \left( \sum_{n=1}^{1007} \dfrac{b^{2n}+1}{b^{4n}+1} \right) < \dfrac{1}{\left(1-a \right) \left( 1-b \right)} = \dfrac{1}{\left(1-|x| \right) \left( 1-|y| \right)}\]

Pheww!! Completed. \(\blacksquare\). – Surya Prakash · 1 year, 9 months ago

Log in to reply

– Saarthak Marathe · 1 year, 9 months ago

Surya, explain the \( \sum_{n=1}^{1007} \dfrac{1}{a^n} \) part.Log in to reply

– Nihar Mahajan · 1 year, 9 months ago

Thank you very much.Log in to reply

@Shivam Jadhav @Saarthak Marathe Please post a solution. My curiosity is running out and so is of some people.Thanks. – Nihar Mahajan · 1 year, 9 months ago

Log in to reply

– Dev Sharma · 1 year, 9 months ago

Agree....Log in to reply

any hint – Dev Sharma · 1 year, 10 months ago

Log in to reply

– Shivam Jadhav · 1 year, 10 months ago

hint is providedLog in to reply

– Dev Sharma · 1 year, 10 months ago

i want to look its solution..Log in to reply

– Saarthak Marathe · 1 year, 10 months ago

Shivam prove the below statement. If you do so, the problem is done.Log in to reply

I found another solution to this problem.This is my own solution. I did not know how to add pdf's to brilliant so I uploaded it on my google drive account. Please See the link below.

https://drive.google.com/file/d/0ByQHFlC74eP_dGxzOTBVaFZKTVE/view?usp=sharing

@Shivam Jadhav @Svatejas Shivakumar @Dev Sharma @Surya Prakash @Nihar Mahajan – Saarthak Marathe · 1 year, 8 months ago

Log in to reply

@Shivam Jadhav Please post the solution!!! I am losing my curiosity!!!! – Svatejas Shivakumar · 1 year, 9 months ago

Log in to reply

– Nihar Mahajan · 1 year, 9 months ago

I think we are unnecessarily wasting our time asking him because he probably doesn't have any solution.Log in to reply

@Nihar Mahajan Yes even I think so. Can we ask Calvin sir to help us in this problem (provided this problem actually exists :D)? – Svatejas Shivakumar · 1 year, 9 months ago

Log in to reply

@Shivam Jadhav Post the solution. – Saarthak Marathe · 1 year, 9 months ago

Log in to reply

@Shivam Jadhav @Saarthak Marathe Please post a solution. Almost three weeks are over since this problem was posted and no one has got a solution so far.

And BTW, in the problem it should be mentioned that x and y \(\neq\) 1. – Svatejas Shivakumar · 1 year, 9 months ago

Log in to reply

– Saarthak Marathe · 1 year, 9 months ago

It is understood.Log in to reply

@Shivam Jadhav Please post a solution. – Nihar Mahajan · 1 year, 9 months ago

Log in to reply

@Shivam Jadhav Please post a solution. – Nihar Mahajan · 1 year, 9 months ago

Log in to reply

– Shivam Jadhav · 1 year, 9 months ago

Sarthak wants to post the solution.Log in to reply

@Shivam Jadhav – Saarthak Marathe · 1 year, 9 months ago

No not yet. Please wait for some more time. I think I am getting it.Log in to reply

– Nihar Mahajan · 1 year, 9 months ago

I think we have got too much time for this. Remember that we have only half an hour in RMO for a question.Log in to reply

– Saarthak Marathe · 1 year, 9 months ago

But once you get such a question by yourself then you will remember the method for your lifetime.Log in to reply

– Nihar Mahajan · 1 year, 9 months ago

Then , please post your solution. Too curious :PLog in to reply

If the following is proved then the problem is done, For \(x\le1\),

Prove that \(\frac{1-{x}^{2n}}{1+{x}^{2n}} \le x \) where \(n \in N \) – Saarthak Marathe · 1 year, 10 months ago

Log in to reply

@Pi Han Goh – Shivam Jadhav · 1 year, 10 months ago

Log in to reply

@Nihar Mahajan @Calvin Lin @Daniel Liu @Joel Tan @Harsh Shrivastava @Xuming Liang @Ammar Fathin Sabili @Satyajit Mohanty – Shivam Jadhav · 1 year, 10 months ago

Log in to reply

@Shivam Jadhav Please refrain from mass tagging. Please try and limit it to five people. Thanks. – Sudeep Salgia · 1 year, 10 months ago

Log in to reply