Difference between revisions of "Cauchy-Schwarz inequality"
m |
m (Adding note to more general form) |
||
(One intermediate revision by the same user not shown) | |||
Line 1: | Line 1: | ||
+ | There are two forms of this inequality: | ||
+ | * {{M|1=\sum^n_{i=1}a_ib_i\le\sqrt{\sum^n_{i=1}a_i^2}\sqrt{\sum^n_{i=1}b_i^2} }} - the common and | ||
+ | * {{MM|1=\vert\langle x,y\rangle\vert\le\Vert x\Vert \Vert y\Vert}} - the rare but more general | ||
+ | {{Todo|More general version http://math.stackexchange.com/questions/1357968/cauchy-schwarz-inequality-proof-but-not-the-usual-one }} | ||
+ | |||
+ | :: {{Highlight|Update: [[Cauchy-Schwarz inequality for inner product spaces]] is a proof of the second form - note that {{M|\Vert x\Vert:\eq\sqrt{\langle x,x\rangle} }} is the [[norm induced by the inner product]] [[User:Alec|Alec]] ([[User talk:Alec|talk]]) 13:04, 4 April 2017 (UTC) }} | ||
==Statement== | ==Statement== | ||
For any <math>a_1,...,a_n,b_1,...,b_n\in\mathbb{R}\ </math> we will have<br/> | For any <math>a_1,...,a_n,b_1,...,b_n\in\mathbb{R}\ </math> we will have<br/> |
Latest revision as of 13:04, 4 April 2017
There are two forms of this inequality:
- ∑ni=1aibi≤√∑ni=1a2i√∑ni=1b2i - the common and
- |⟨x,y⟩|≤∥x∥∥y∥ - the rare but more general
TODO: More general version http://math.stackexchange.com/questions/1357968/cauchy-schwarz-inequality-proof-but-not-the-usual-one
- Update: Cauchy-Schwarz inequality for inner product spaces is a proof of the second form - note that ∥x∥:=√⟨x,x⟩ is the norm induced by the inner product Alec (talk) 13:04, 4 April 2017 (UTC)
Contents
[hide]Statement
For any a1,...,an,b1,...,bn∈R we will have
n∑i=1aibi≤√n∑i=1a2i√n∑i=1b2i
Proof
Basis for argument
Consider first the function f:R→R give by f(x)=ax2+bx+c
If f(x)≥0 then using the quadratic equation we know the solutions (to f(x)=0) will at be: x=−b±√b2−4ac2a
As we want f(x)≥0 we must have either a repeated solution (a point where f(x)=0) or no real solutions.
In the first case (repeated solutions) we require b2−4ac=0 as then −b±√b2−4ac2a=−b±02a=−b2a - our 2 repeated solutions.
In the second case we require b2−4ac<0 as then the √b2−4ac term will be imaginary, thus giving us no real solutions.
Conclusion of first argument
We conclude from this that if a quadratic ax2+bx+c is to be ≥0 then b2−4ac≤0
Core of argument
In the basis we required a function, f(x), we will now build this.
Take n∑i=1(ait+bi)2 and notice:
- n∑i=1(ait+bi)2=n∑i=1(a2it2+2taibi+b2i)=t2n∑i=1a2i+2tn∑i=1aibi+n∑i=1b2i - which is a quadratic in t
- ∀ai,bi,t∈R (ait+bi)2≥0, so n∑i=1(ait+bi)2≥0 - our quadratic in t is ≥0
Using the above this means b2−4ac≤0, where:
- a=n∑i=1a2i
- b=2n∑i=1aibi
- c=n∑i=1b2i
Conclusion of argument
4(n∑i=1aibi)2−4(n∑i=1a2i)(n∑i=1b2i)≤0⟺(n∑i=1aibi)2≤(n∑i=1a2i)(n∑i=1b2i)⟺|n∑i=1aibi|≤√n∑i=1a2i√n∑i=1b2i
But as x≤|x| (recall |⋅| denotes absolute value) we see:
⟺n∑i=1aibi≤√n∑i=1a2i√n∑i=1b2i
QED