# Teaser Inequality

Here's a "teaser" inequality for an upcoming article I'm writing on an inequality I found that has this as an application:

Given that $$a_1,a_2,\ldots a_n\ge 1$$ are reals then prove that $(a_1^2-a_1+a_2)(a_2^2-a_2+a_3)\cdots (a_n^2-a_n+a_1)\ge a_1^2a_2^2\cdots a_n^2$

For now, I wish to see solutions with inequalities we currently have. Good luck :3

copy-pasted from AoPS lol

Note by Daniel Liu
3 years, 8 months ago

MarkdownAppears as
*italics* or _italics_ italics
**bold** or __bold__ bold

- bulleted
- list

• bulleted
• list

1. numbered
2. list

1. numbered
2. list
Note: you must add a full line of space before and after lists for them to show up correctly
paragraph 1

paragraph 2

paragraph 1

paragraph 2

> This is a quote
This is a quote
# I indented these lines
# 4 spaces, and now they show
# up as a code block.

print "hello world"
# I indented these lines
# 4 spaces, and now they show
# up as a code block.

print "hello world"
MathAppears as
Remember to wrap math in $$...$$ or $...$ to ensure proper formatting.
2 \times 3 $$2 \times 3$$
2^{34} $$2^{34}$$
a_{i-1} $$a_{i-1}$$
\frac{2}{3} $$\frac{2}{3}$$
\sqrt{2} $$\sqrt{2}$$
\sum_{i=1}^3 $$\sum_{i=1}^3$$
\sin \theta $$\sin \theta$$
\boxed{123} $$\boxed{123}$$

Sort by:

I'm going to bed now, so maybe I'll try this later. I's just wondering if the following manipulations help!

Define $$b_i=a_i-1$$ for all $$i$$. Then we gotta prove that

$\prod_{\text{cyc}}\left(a_1^2-a_1+a_2\right)=\prod_{\text{cyc}}\left(b_1^2+b_1+1+b_2\right)\ge\prod_{\text{cyc}}\left(b_1^2+b_1+1+b_1\right).$

Can anyone finish this from here?

- 3 years, 8 months ago

nice observation. for the n=2 case we can directly use C-S to prove it. it doewnt apply to higher cases though

- 3 years, 8 months ago

Great observation. Which inequality allows you to justify that step?

Staff - 3 years, 8 months ago

Hi, just wondering. How do you save stuff to sets? I would press create new note, but it wouldn't save to the set. What am I doing wrong?

- 3 years, 7 months ago

@Calvin Lin I can't seem to solve this using Induction by the straightforward way... Can you try this out?

- 3 years, 8 months ago

Induction would not an approach that I would think of, mainly because the "cross terms" do not result in anything nice.

If we insist on trying that, the straightforward way requires showing that

$\frac{ ( a_n^2 - a_n + a_{n+1} ) (a_{n+1}^2 - a_{n+1} + a_1 ) } { (a_n^2 - a_n + a_1 ) } \geq a_{n+1} ^2.$

This is equal to

$( a_1 - a_{n+1} ) ( a_n ^2 - a_n - a_{n+1}^2 + a_{n+1} ) ( a_n^2 - a_n + a_1 ) > 0$

which is not necessarily true.

Staff - 3 years, 8 months ago

A simple induction will do it, I guess

- 3 years, 8 months ago

Why don't you try the induction? You may be surprised.

- 3 years, 8 months ago

Every set of n numbers can be arranged in ascending order.Let $$A_n$$ be the LHS of the inequality and let $$a_1\leq a_2 ... \leq a_n$$.For n=1 the inequality is obvious.To make the change from n to n+1, we see that $$A_(n+1)=\frac{A_n}{a_n^2-a_n+a_1}.(a_n^2-a_n+a_{n+1}).(a_{n+1}^2-a_{n+1}+a_1)$$,so we want $$\frac{(a_n^2-a_n+a_{n+1})}{a_n^2-a_n+a_1}.(a_{n+1}^2-a_{n+1}+a_1)\geq a_{n+1}^2$$ .Let $$x=a_{n+1},y=a_n and z=a_1,x\geq y\geq z$$.Then we want to prove that $$(1+\frac{x-z}{y^2-y+z}).(x^2-x+z) \geq x^2$$. That is equivalent to $$x^2-x+z+\frac{x-z}{y^2-y+z}.(x^2-x+z) \geq x^2$$.Now we can cancel x^2 and factor out x-z: $$(x-z)(\frac{x^2-x+z}{y^2-y+z}-1) \geq 0$$, which is obvious since $$x\geq y \geq z$$.

Q.E.D

Is this solution correct?And how can we prove the inequality without induction?

EDIT:I've only proved the case when the reals are in ascending order.But isn't the value of the LHS minimal when this is the case?

- 3 years, 8 months ago

I would have issues with assuming that they must have a certain order. It is (as yet) not obvious that this minimizes the LHS.

Staff - 3 years, 8 months ago

Yes, that is the main problem of my solution.I don't know if we can even fill the hole

- 3 years, 8 months ago