hardest topic (1 Viewer)

McLake

The Perfect Nerd
Joined
Aug 14, 2002
Messages
4,187
Location
The Shire
Gender
Male
HSC
2002
Formula:

n /(1/a1 + 1/a2 + 1/a3 + ..... 1/an)

Explanation:

n is divided by the sum of fractions to n
[In the case where k = -1, for a HM]
 

spice girl

magic mirror
Joined
Aug 10, 2002
Messages
785
An example? try this:

Given a1 + a2 + ... + an = k, and a1, a2, ..., an > 0

Prove 1/(a1) + 1/(a2) + ... + 1/(an) >= (n^2)/k
 

spice girl

magic mirror
Joined
Aug 10, 2002
Messages
785
You just reverse the process and answers become questions...it's not hard to make these things up...

and beside, these AM-whateverM inequalities often begin like: if a1 + a2 + ... + an = something, then blablabla..
 
Last edited:

Weisy

the evenstar
Joined
Aug 2, 2002
Messages
656
Location
Here
Gender
Female
HSC
2002
I like all this stuff, it's fascinating

but do we have to know harmonic mean at all for the HSC? It's only that it was never even mentioned to us.

Also in the example that spice girl gave, n is integral?

It would be great to see the proof of that example...I've tried, except I don't think my working's legal. And I've also never done a problem like it before.
 

BlackJack

Vertigo!
Joined
Sep 24, 2002
Messages
1,230
Location
15 m above the pavement
Gender
Male
HSC
2002
Yeah, it's integral...:D n is the number of terms 'a'.

I haven't heard of HM before either. I wonder what it means... I don't think we'd need this in the HSC though, harmonic means aren't mentioned in the syllabus... (however, if they lead us up to a proof...)
 

spice girl

magic mirror
Joined
Aug 10, 2002
Messages
785
Knowing the harmonic mean isn't necessary, but in 4umaths, using any method is legal. Personally i've never used the harmonic mean crap, there's usually another way. But anyways:

a1 + a2 + ... + an = k

(a1 + a2 + ... + an)/n = k/n

You look at the expression: 1/(a1) + 1/(a2) + ... + 1/(an) and you recognise it as part of an HM expression. So you use AM >= HM:

(a1 + a2 + ... + an)/n >= n / (1/(a1) + 1/(a2) + ... + 1/(an))

k/n >= n / (1/(a1) + 1/(a2) + ... + 1/(an))
thus 1/(a1) + 1/(a2) + ... + 1/(an) >= (n^2)/k
 

Users Who Are Viewing This Thread (Users: 0, Guests: 1)

Top