UCSD Mathematics | Home - Determinantstkemp/18/Math18-Lecture19-C00-after.pdfProperties of...
Transcript of UCSD Mathematics | Home - Determinantstkemp/18/Math18-Lecture19-C00-after.pdfProperties of...
Today : § 3.3 Determinants
&§5.1 : Eigen values
Next : § 5.2 : Characteristic Polynomial
Reminders
MATLAB Assignment # 4.Due TONIGHT by 11:59pm
My Math Lab Homework # 6 : Due Tuesday by ltisepm
Midterm # 2 : Next Wednesday,
8- 10pmRom 1 seat assgn .
& practicemidterms posted on course webpage .
Properties of Determinants
* Defined by cofactor expansion ( recursive expression , expandalong any now or column )
* Behaves nicely under new ops : detA= ± ( product of pivots inrow reduction Kerrey ) ) .
* detA=o iff A is net Invertible* det : Mn → IR is multi linear in each now and column of
the matrix,
with the others held fixed .
* dot LAB) = detlA)deHB ).
.: detl An)=(deHA)Y .
Two MORE PROPERTIES :
det l At ) = Ydef Adot ( cadb )
= ad - be
"
det l At ) = det (A) debtgcd)
= ad - Cb
What is the determinant ? What does it mean ?
Parallelograms :
b.b.to
S
A IQ'
g
Area Plaobj + Area Pla. ,s| = Area P[ a
,
bts ]
: the R is multi linear
d La ,b I + d [ a.,
s ) = d ( a. b.ts ).
heorem . Given two vectors g. b. e- 1122,
the area ofthe parallelogram they determine B
Ala. ,
b.) = ldetla,b.)
lydettdettt" '* "
tutaidwhat dees the sign of detlg
,bt mean ?
orientation.
What about nxn determinants with n > 2 ?
Parallelepipeds taped
p Idetlabcjfldumelp )
§ 5.1 / A linear transformation T :v→V tends to meue
vectors around.
Eg . TIM
.tn#=l3oylgsTl!2)=sojogenvawe412,1t.ly/=ygyoe9enEK
Eg .
If A is a stochastic matrix ( new sums =| ) thene%Yt↳
De¥t¥¥tgIdH=tgt¥eh¥¥M.Hi .tt#- eigenvector
Definition Lett :V→V be a linear transformation .
If there B a non - zero I← V such that
[email protected] v. an
eigenvector of T.
The scalar X is
the corresponding eigen value.
( We similarly
talk about eigenvectors trgenvalnes ofsquare
matrices.)
Eg.
Show that 7 is an eigenvalueof A =L 's{ftp.fyjffpg)
Ie.
there 's Its with AI ' 7k.
n =7H )i.e. e. = A .x
- Zx = (A - 7I ) I want
a # Halted 'f8fH 's:H→HHattabaugh
Theorem : Forany he R
,
the set of vectors v. for whrh
TH ) = tv
B a subspace of V,
called the eigenspace of X.( ( We only call d an eyen value if its eigenspale B He})
{ ± : Tkthv } = Nul ( T . JI ) Is a subspace .
Hit could be {e. } . If so
,X
B ne± an eiyenuahre .
Eg.
The rotation matrix ( o
, f) kcwsi ) has no
eizenvahei.
If d were an egenvahe ,
then the is an egnvectr HIT feel
"
t¥Y¥:*,µ .:¥%Y¥a-
y .
. Hy i. HIDY- o
> o; .y=@(¥:@
* Given a ( potential ) eigen value,
finding the corresponding
eigenspaee is noutme.
* Finding the eigen values is typically hard. ( Row reduction
changes the eigenvakes . )