How many proofs of the Cauchy-Schwarz inequality are there? Is there some kind of reference that lists all of these proofs?


$ ewcommandbx<1>,box<15px,border:1px groove navy>displaystyle#1, ewcommandislovenija-expo2000.comrmi ewcommand ext<1>slovenija-expo2000.comrm#1 ewcommand oot<2><>^#2sqrt<#1> ewcommandderivative<3>fracslovenija-expo2000.comrmd^#1 #2slovenija-expo2000.comrmd #3^#1 ewcommandabs<1>leftvert,#1, ightvert ewcommandx<0> imes ewcommandsumm<3>sum^#2_#1#3 ewcommands<0>space ewcommandi<0>slovenija-expo2000.comrmi ewcommandkume<1>slovenija-expo2000.combb#1 ewcommandold<1> extbf#1 ewcommanditalic<1> extit#1 ewcommandkumedigerBETA<1> m #1!#1$

Here"s a simple proof:

$|vecxcdotvecy| leq |vecx||vecy| $

Substitute $|vecxcdotvecy| = |vecx||vecy|cos heta$

$| |vecx||vecy|cos heta |leq |vecx||vecy| $

Divide both sides by $|vecx||vecy|$

$ | cos heta| leq 1$

-Hey, I was looking for a "more serious" proof!

Then here you are!

Here"s another simple proof:

This is projecting a vector to lớn another one (Click the gif if it doesn"t load):

*

You drag its end in a line that is perpendicular to the other vector. Then multiply the length of the new vector with the old vector.

Bạn đang xem: Real analysis

Do you know what the multiplication is equal to? The dot sản phẩm of the vectors

*

When you project that vector, its norm (length) becomes lower - or stays the same if one of them is a scalar multiple of the other one.

^^ That was the proof. Think about it.

Source: $3$Blue$1$Brown

Wait, I look for a "really serious" proof!

Here you are.

Another proof:

Let $p(t)=||tvecy-vecx||^2$

As there"s an absolute value, it must be equal khổng lồ or bigger than $0$.

$p(t)=||tvecy-vecx||^2geq 0$

$p(t)=(tvecy-vecx)(tvecy-vecx)geq 0$

$p(t)=t^2(vecycdot vecy)-2t(vecxcdotvecy)+vecxcdot vecxgeq0$

Let"s substitute some things.

Xem thêm: Sách Tài Chính Doanh Nghiệp Hiện Đại Trần Ngọc Thơ Pdf, Tài Chính Doanh Nghiệp Hiện Đại Trần Ngọc Thơ Pdf

$p(t)=t^2underbrace(vecycdot vecy)_colorbluelarge a+tunderbrace(-2vecxcdotvecy)_colorredlarge b+underbrace(vecxcdot vecx)_colorgreenlarge cgeq0$

$p(t)=colorblueat^2+colorredbt+colorgreencgeq0$

Its minimum value must be $large frac-colorredb2colorbluea$

Substituting $large t= frac-colorredb2colorbluea$

$p(frac-colorredb2colorbluea)=colorbluea(frac-colorredb2colorbluea)^2+colorredb(frac-colorredb2colorbluea)+colorgreencgeq0$

$p(frac-colorredb2colorbluea)=colorbluea(fraccolorredb^24colorbluea^2)+colorredb(frac-colorredb2colorbluea)+colorgreencgeq0$

$p(frac-colorredb2colorbluea)=fraccolorredb^24colorbluea+frac-colorredb^22colorbluea+colorgreencgeq0$

Forget the $large p(t)$ function side (LHS)

$fraccolorredb^24colorbluea+frac-colorredb^22colorbluea+colorgreencgeq0$

Multiply by $large 4colorbluea$

$colorredb^2-2colorredb^2+4colorblueacolorgreencgeq0$

$-colorredb^2+4colorblueacolorgreencgeq0$

$4colorblueacolorgreencgeq colorredb^2$

De-substitute

$p(t)=t^2underbrace(vecycdot vecy)_colorbluelarge a+tunderbrace(-2vecxcdotvecy)_colorredlarge b+underbrace(vecxcdot vecx)_colorgreenlarge cgeq0$

$4colorblue(vecycdot vecy)colorgreen(vecxcdot vecx)geq colorred(-2vecxcdotvecy)^2$

Using the identity $large vecvcdotvecv=||vecv||^2$

$4colorbluecolorgreengeq colorred(-2vecxcdotvecy)^2$

Using the identity $(f(x))^2=(|f(x)|)^2$ (where $f(x)inkumeR$)

$4colorbluecolorgreengeq colorred-2vecxcdotvecy^2$

As the both sides are not negative, you can square root both sides.

$2colorbluecolorgreengeq colorred$

$2colorbluecolorgreengeq colorredvecxcdotvecy$

$largecolorbluecolorgreengeq colorred$