Today 6.1/6.7 Products - UCSD Mathematics | Hometkemp/18/Math18-Lecture23-B00-after.pdf · 2018. 3....
Transcript of Today 6.1/6.7 Products - UCSD Mathematics | Hometkemp/18/Math18-Lecture23-B00-after.pdf · 2018. 3....
Today § 6.1/6.7 : Inner Products
Next : § 6.2 : Orthogonal . y
Reminders
My Math Lab Homework # 7 : Due THURSDAY by ltispm
MATLAB Homework # 5 : Due FRIDAY by 11:59pm
Final Exam is on Saturday,
March '7.ly?foapmm
@ the BEGINNING of Exam.
Week.
An nxn matrix A is diagonal sizable if A =
PBPtenvertiblT
diagonal
Thisis equivalent to Rn having a basis of eigenvectors for A.
This happens , e.g.,
wkw A has n distinct real eigen values.
But it cay fail .
↳ Eigen values : roots of PACD = detllt - XI ) ← degree n polynomialmust have all real roots for Ato be diagonal izable
.
↳ The algebraic multiplicity of X is the degree of this root in p*The geometric multiplicityof X is dim . Nul ( A- XI )
.
Always havegeom.mu It A) falg.mult.ly an be §
A= Hosey Bs HeIII. Pa
'tPBMaIg§¥f!YYsz
dim Nul ( A - SI ) = 2,
dim Null B- SI ) . -1.
" 4 ' s ±
Theorem If A has a repeated eigenvatue X,
then
dim Nul ( A- II ) s multiplicity of X.
The matrix is diagonal izable if and only if
PA has only real roots,
and dim dull A- II )= multiplicity of X for each X
.
In this case , any bases for the eigenspaces
together form a basis for R"
.
So,
just whizb matrices are diagonal izable ?That's a really tough question to answer
.
A partial answer
will be given at the end of the course.
To get there.
..
§ 6.1 Definition : The dot product or inner producton 112
"
is defined by it nxn
m -
y . .V= < b ,y > = u.tv = yty ¥ butnxl hxl lxh NX ) ↳ nxllxw
ftp./Ygf=luikujuH=uii+kv+usg .
Properties:c
, a.v. =±u .
fold(2) ( y,
+ b ) . .v= bit tuz . I }tissaysif {= uu ;D
1- ( g)' - g.v. We,±u.y±= l.u.tt#v=.nPv+ufIvT : R "→1R is
almewtrarsfomatn (3) ¢y) . .v
(4) b.y= up +422+45>0 unless be, hwkhjytf
.
The length of a vector is defined t.be
11×11 = VIMore generally
,the distance between two vectors y
,!
13 distly,
!)= Hu.
- tll x,
E.g. a. =H| ,
tffl:
2.
,- - - - . . .
,
t ;
drstlu .u=Hb-±H=H # H -
ax .
/ / l / Xz-
.
y
=jM4C -25+1-25.
=jTt4t=r=3.
.
X,
How does11A. v. 11 relate to Hull,
11111 ?
:-1112 = ( b- I, y- v. )
= < ±,u . ± ) + s - ± ,u.
. x > -
=ky,u.
) +4¥'+t±u .> +2¥ - t >
= < is ↳ > - < a ,v.
> - < %u. > + tiny >
= 116112 - ZCY,
.V > + 111112
i. 114112+111112--1141112+2<4,
! >
If < apses then utrrectsntem to Pythagoras.
NYH 4118112 . My. V. 112
*If it's o
,we get JTkorm
,✓
u.tv applies to rghtagks .
Law of cosines In 2 dimensions ) :
bC
a2tb2=c2+2abces€a) a=HbH b=lKH c=Hy .gl
" (nylptnvlp .tn#p+2nulllKll6soat
; Yas " n' + " IH'
' Hy-±Nt2( y,
, >
4< y, .v ) = 1141111.41656 .
< a ,v.
>Gsb E Ebl ] 8=65'( pay , )
o÷vs-
-
-
Cauchy - Schwarz Inequalityt
Riesz ly. v. I s 1141111111.
Hike → Segal → Gross → Kemp
Pf. fct )= in-tvlp 30
= s a- t.v.u.tv > = < u. ,u,
> - tcu ,D - Kyu.
>
+
ttkv.us#.:meIy*Et*" it
nots :- b±IbT4ao .
:b'
. 4ac.co
← Hu,±>Tf 411411411112