# Abstract Inequalities

1. Oct 27, 2012

### odolwa99

Can anyone help me confirm if I've solved this correctly?

Many thanks.

1. The problem statement, all variables and given/known data

Prove that $\sqrt{ab}>\frac{2ab}{a+b}$ if a & b are positive & unequal.

2. Relevant equations

3. The attempt at a solution

if $(\sqrt{ab})^2>(\frac{2ab}{a+b})^2$
if $ab>\frac{4a^2b^2}{(a+b)^2}$
if $ab(a^2+2ab+b^2)>4a^2b^2$
if $a^3b+2a^2b^2+ab^3-4a^2b^2>0$
if $a^3b-2a^2b^2+ab^3>0$
if $(a^2b-ab^2)>0$...true

2. Oct 27, 2012

### Staff: Mentor

You are essentially starting off by assuming what you are to show.
Why? There's no guarantee that a2b - ab2 > 0, so you can't say with any certainty that this is true.

Also, you should not be starting each statement with "if". The hypothesis (the "if" part, the part that you assume to be true) for this problem is this:
If a > 0 and b > 0 and a ≠ b

The conclusion (and what you need to prove) is this:
Then $\sqrt{ab}>\frac{2ab}{a+b}$

Each statement that you write should imply the truth of the statement that follows. The symbol to use is $\Rightarrow$.

One way to do this problem is a proof by contradiction. Assume that a > 0 and b > 0 and a ≠ b, AND that $\sqrt{ab} \leq \frac{2ab}{a+b}$

When you arrive at a contradiction (and you should), you will have proved the original statement.

3. Oct 27, 2012

### odolwa99

I can see what you're getting at, but I've included an image of an example question from the text book I'm using. Is the books approach wrong, or have I followed their method incorrectly?

#### Attached Files:

• ###### math (2).JPG
File size:
70 KB
Views:
135
4. Oct 27, 2012

### aralbrec

I don't like that notation either. Implies is much better.

The very last step is not right. Trying factoring out ab and then notice you have a quadratic equation that can be factored. The proposition is false if you can find *any* a,b satisfying the conditions to make the inequality untrue.

5. Oct 27, 2012

### odolwa99

Ok, thank you for checking that.

6. Oct 27, 2012

### SammyS

Staff Emeritus
$a^3b-2a^2b^2+ab^3\ne (a^2b-ab^2)$

Factor out ab from a3b-2a2b2+ab3 and you will get ab times a perfect square.

Regarding you textbook's example, which I think we've seen before:

It reminds me of the steps one might take on scratch paper to figure out how you might go about doing the actual proof. Starting with what you want to prove, you work backwards toward something(s) you know to be true.

So, for the case of the example:
Assuming a>0 and b>0,

$\displaystyle \frac{a}{b}+\frac{b}{a}\ge2\$ is true if $\displaystyle a^2+b^2\ge2ab\ .$

$\displaystyle a^2+b^2\ge2ab\$ is true if $\displaystyle a^2-2ab+b^2\ge0\ .$

$\displaystyle a^2-2ab+b^2\ge0\$ is equivalent to $\displaystyle (a-b)^2\ge0\$ which true.​

Most mathematicians would look at this proof as being written in reverse of the usual order.

IMO: The following is a more conventional order.
Assume a>0 and b>0.

$\displaystyle (a-b)^2\ge0\,,\$ the square of any real number is non-negative.

Expanding the left-hand side gives us: $\displaystyle a^2-2ab+b^2\ge0\ .$

Adding 2ab to both sides results in: $\displaystyle a^2+b^2\ge2ab\ .$

ab>0, so dividing both sides of the inequality by ab gives the desired result, namely: $\displaystyle \frac{a}{b}+\frac{b}{a}\ge2\ .$​

7. Oct 27, 2012

### odolwa99

Thanks again, Sammy.

I was on the this site, a few weeks back, with questions about Proof by Induction, and there was a lot of disagreement then about how those questions should be answered. This book is actually meant for state exams, so I don't know if a drop in the quality of solution, compared to university level math, is a consequence of that, or just a matter of preference by the author of this particular text...?

Either way, thank you for the clarification.

8. Oct 27, 2012

### aralbrec

The notation in the book is not incorrect; it's just not what most of us are used to seeing.

If..then is the logical statement implies.

If A then B

is the same as

A -> B

The statement is false only if A can be found to be true while B is false. This is exactly what the problem in the book is doing, based on a chain of "if A then B, if B then C,..."

9. Oct 27, 2012

### odolwa99

I'll make note of this, too. Thanks.