Chapte-4

Valuek and vector,

agnaltzdtion Eigen eisn
squaie mcn'a. The matnx A
Defnition het A be an n-
nan-Singas
Sad to be diagonalizable ftheve Ckist a

maix P s.t.

PAP B is a diagonal matna

cas. hancfmaon Point of Vieo

T V V
L.T if
hexe
theve exisk
exish aq baalb
baato
be eliaanalizable
Then T
Then is saldt

S of v s.t 9iagonal mati'

C1 ur-t. S)
man repesentahon ot T

squialerce o the Two koin View ., T definad by

viewed. a
A manx A4 can be
A
TOx) =AX
whasX ER".
Thun CJe A
-standad baut E t S.
S.
on basi
et P= Tsansiton maix S.
the s
Coum a he eenerts in

vepwentahion T Telahve to, bduts S

Then
hen P P AP =
mahk new

Kole of Uncleylying fiel k: The diaganalz-ahon amatix A/T

will depend on the o of apno'al oves tid! k
c tH-o. Hhas ho 300T k=R,
 olpomals of mamces
ove fiekdk
flel-4,t+ t "
+ - -

+ 4,t + 4,
d A E Mn - 89uale mdhk.
d
Then fA) A+ a,
n A+ -.
+A 4
Tderdtihy madhx
A s a root of ft) fA)=0, zeo mai'

t ) 2t-3t + 5

2
Find fA)= 2434 +5L

=7 o
3 4J34 22J
[16 147
21 37

wh glt) =
t2 5t-2
Calaulat qA)
52 +-s -f1o7
A5A 21 =|7 1o
A)=
-

o o

8lA )=0
A is a soot of q(t),

Mars and Lineas Caairs Ad T: V->V be .T.

T ToT TTOT= TOTT
HT= 9,T+T +T+f
Tdentity mappng.
 T is fT)=0, the zn0 mappng
a zero Toot of ft)
mahix A
het fu and qe) be polyrcmials Fr any uare
1 4.
.

and calas R,

v (t+f) A =
ftA)+gUA)

Ci) kUA) k flA)

0
O) fA g{A) =gAHA) 4,t + ao
het fit- 4t+ t
h t + h%
Gltl
it i > m.
t m n and het b o
+4tb,)t + (ot
(4t h) tH4
() a,t-tn) t+
-
-

f1gA (a,th)A
.

=4 a,A
bnA+ bA+
-
-

+h Atbo

fIA)+ 2(R
Dtm
nth
+ t +
+ m t

ht Cnt
+a, bm-14,n +4,&
Dtm4 47,b

nm
U, m-2 bm7
tn-2
C
 ntm
tacAl-GA
n

toigonl-(ar- / k=o

Charactesshic olynomial Cayley -Hamilton heaen

n- squaie maik.
het A= Laj be a
is a vaiable
t
M A t Ln
n-squa mamx
n

Thendf-det (-M det (t ,-A)

()"dt (A -tI» 3t ic aa

is cald the chaactevishe polyramial of A.

olymomial of degsee
5

Exnple i) A= 4 20
(-1)(t-2)
7-A-&-a -s|-
-

-4
=-3t+ 2-20
At-t zt -18 characteichic polynomial of

) c [ -3
+/8
at-ltr-Al= |-3 =(t t3)(t-3)
t+3
t) =t +q s
chauac)aste lgmial ofC
Example Find the chavacteushe polyromial o
A- 2 3
32) 4
6 5
Altle |tI- Al= | t- -3

-3 t -y
- 6 - 4 t-5

+ 6t- 35* -38

V.Tme
Therem92 Evey m aix Als a root of
ECayley-Hamiltm)
(Only stale met) chara cteishic palynomia.

Le. AlA) = O.

A At-|tI- Al
Chakactenistrc polynomial: = t6t - 7
= |t-1 -3

-4 t-5

4lt1=A 6A -7I
A343J-|2 19

6 18
13
[ Lo 1
4(A) =

24 37
37 24 30
1-6-7 19-18-O|-(o
24-21o 31-30-71
i a roo of A(E.
 2
Ex A
3

-2

AE= - A l =t- t-3 2

t37 B1-17
3 t-9

A and 8
are
Two mantes
Definihion: (Similas mabites) hon-6ihgula
be simula mahtes if 3a
Said to
mamxP s.t
B-PAP
have the 8ame
Chavactesichc polynanil.
Similas maies
Themem
8- PAP
t
of A- |fI- A|
chavadeuste polynomial
4 (t 6-+I-8
4ltl=

4 1e)-det(tr-B) T= P)
de (6- FAP)
=

17 P )
ddr tP-F'h
d[P(4I -AJP] det(Ae)=dd)detB))
is ABl-IAJJB/
det(P) dd (tI-A) detle) PP = I
IPllp/=/-/
1 . det (tI-A)
 sace of a marix Cofactoss
di3
A

TA)= ace oA =
4 +Q+asa = Sum of diagnal entries

a =-) q3 =/a2 Q3
Co o-fods o
43233 Ca3
C-fadmo 9, * | 4 ay3 = |a as
43

3 5
A 6 12 0

14 15

mamces oOmdes 2 and3

characdeiste_folynamiad
Fosmulas fo

t- (tu/Al) t 4 ddlt)
(ayt0a)+dtlt)
=

A)- t2

Aa+ Ass) t dortA)

A)
tylA)t2 + (Ant
4)- t
Cofactrs o , 422,a3 pechels .
A, Aza, Asa
S htA), S At Aat Az3 S detlA)
Kemayk
A t - t - s , t*+ S t - Sa
S= Sum of al nincipelmins
3, Af odisk
 n-squak madh'x Then i
Theorem-44 Let t b e an

chanaceistic palynomal is

S,e+ S,t +4tPs,

R)= t
de.k
Whese S, is the Sug othe pincpal min s o

At-e(t -
E x a m a l e

22
-S
6
2

- 5 -

At)=t-2
-
t-4 2

O t-6 5
- th3
= (t-2t-4-2 -? t~s
- -

t-6 5 6t-6 5
o -2 t+3 -2 tt3

=(a)t= (t-)t+3}410}+ t-/t+ 3) +io

-t-o(t+3)lo-9)-s]
=(t-3-8)( t-6t+3
-3-c)t+(-8 +3+18)t + (4s-9)t -2
Al = t"-4t°+13t*+34t -2
Ty , t + St-s,t t Sy
S
tlA)-= 4+ a +a33 7y = 2t4t6+-3) =
9
S
Su allnapgmingaof oykis
 ,a C1,3) (,4) (2,), (2,y,3,4)

= 18+1o) +(-11) + 24- 6 +12 +(8-9

8 - a + 24-6 +6+3 2t 5-2 =3
5 2 / + / 2 5s

S=4 6 5
o6 s o 6
-3
6 O -3

+ 2(-12)--15) + 6(3
= 4-19+I o) +2 (-/8+lo)
- 3 2 - 1 6 - 2 y +5+)8
-72+33 =-39

dt/A) =
- 2
Sy
ACt) = t 9t3+13t+ s9t -24

T
CharacteNstc olynomial of a Hneas opesat
het T:V V V be a L.T and dimV fnt
Then the tharacteaishic olynomias fT is defined a the
the
choncdeicthc polyromias any man sepsesentaton oT

Let A and be watn'x sepoesentations o T ur.t baets

B
S and Sespechvely Then
maik
B AAP whexe P is the change-of-baais
Tem S to S
A and B are simikh mace
Aand B have same charadeishe alynomial

T e chovaderishe polynminl of T is indepenen

The the
kashulas kaais in whrh the ma representatien of 7T

5 Computeo.
heade 9.2 Cayley- Hamiton A opeAaas| lureas onsbnah
lineas
Tis a zexo of icharacteistic polynomial.

eigen alues and egen Vecto

iagonalezahon,
A
A can repraceniel by
be repraenil
Can be
matnx. Then
A be n-squake
t
CSimias to any
a diagonal mahia D=| a AkoglA,b"ud)
n
C
thexp
thewe exisk baus S conaisting ovecterk 4,42, 4
st Auy= 4
Au -
Aun- Ann
o o-
A4 U .

Un[4 4 un ]
w m m m n

Baus vectos

dhgral arix D
AP PD
D- P AP

matix
Then a &calas A
het A be any Bquae
Derinihon:
callhd an eigenvalue of Aif there kisk
mon-zem Ccolumn) vec s.t
At t -

vector sahifing thi celation 1 iis called an

Any
cigenvectos A belonging To the eigenvalue
 i s an eigonectos belonging to eigen value A and
Note
Rs eny scalas, then ko) f's albo ah eigenvector belagingy

4o .
A= a O
AChu) =
klAu) A(au
use
Ake) a(ke)
A* = du* uLke is an eigenvecta
also
belonaing eiqenalue A.

to
Note 2 The set E =
Collecicn of all eigenecoss belanging
cigon aHe A is Subspace o V.
alue

=Eignspace of
callud an eigenlne and
sf dinEl, then Ea s
is callda scaling facter
Eigen values
tharactevishic value rape Value
Vedos Vecta Vec

mecnx is imias to a diagonol

hecie 9.5 An n-&quare

mechixi'ff A has n lineasly independlent eiger- Vactok

eumentz o D are he coneapomeiney
this case, he diagonal
where P is he matnx whase
enealues and D= AP,
Colum ns qve the cigenvectors.

bxample (cpen klues &igenvecta

A-3-y
2 -6
tI-At-3

aChasadesishe kolynomiadefA: -2 t+6

 At) = (t+3t -1o) = t-2)/t t5)
t : 2,t -5
Rootcharacdesiche polynomial
ase

values A ae 29 - 5 |
So,he eigen
Hsen vector 2
Ax 2 X

FA-21)X =O
227
3-4
L8
3-uy 2 -4y-0
27-64 20

2 vaiables, i equahin -4y=o

Tee vaniable, | >pivot vasiable.
4-t, *= 4t het t=/
het to =2,
contaponing
X=(4,1)'> cigen vecta
ThenX
5
Solurian space e homo, Sstem
-5
Egen ector a=
ATX=0 (A+5T)Y =
8-4y0
d
2 2 o

27--0
t t, 2t

genvect = (1,2)

penapace=-5 is Span31,2)f
 then 3 P st
f A is diagonalizable,
AP D =diag(Fis Ra,, kn

A PDP
(PDP-) (PDp)
= P`P
A
A A (P* )¢Dr)
=

ASPD° Pr
T Genenal
-Pp -P[4igl}, P
Then
t f) e any olyromias.
) = t * + qt +9
fUA)= %A+4A + 4I
(PDP)+ 4 I
4(PBP")+ q
fPDP) =

ue
4(PD Pr)+ 4(Pp)
(PD D-) +

=P4D+4D+41 P
PDP)= PflwP 6
tPDP) = P
dig ( fR), flk), ---, flk)Pr
Dae Ldt
non-negative, Lt
dngoral enmie o
B P clngWki, Vka, -"NEP Ce. BA, And
sot dA
nonugaive 8qua%e
Ihen B is a
nnnagaive,
the eigen values o B qse
Examble9S

A31-7-L1 o
H =
1.U , is an eigenector o belorging
belorging to
ciqen-value A=1

:]]-"] A elaging to
i s cn egenudts o
Au =y eigervaluue d2=4.

Heo, , »i =3#o
, a L.I.

Matniy A is diagmaizsble.
-/3
P- , P P
O

dngl, 4)
-2

85

APer- sJ2sá
86
 het ft= t st*+ 3t +6.
flAa)=fI4) = 2,
HAN-ftu) =5,
3
fA-PfDe =[ Ts

asitihe Squas Rod of A A=l, Va = 2.

-3
B PD P =

V3
B =
B A ( Check)
-in
Remah P = - |in

&gen-Values cind Een ecbs

upeties aof then
hes he
h e ollowing
falawing
anl

squas mam*,
mank
A be a
Theorem 9.6
equivalet
an eigen
value A,
a is
i) A Lalas.
chacactesistie olynomial o A
calar A is a oot of
di) The
ce. A l ) = 0
singulas.
M= A -AI
) The m x
MX=0.
homogrunu system MX=0
pace the
Salution
fams
Eigenspace E Collecttan o all 8uch X
Ax A X>
AX- AX =O
- ) X =
0
Homagentous system with|M=
A-AI
1
olin
ths homogennaus system has infinitely many
Fomi),
 SOme maices have no eiqenvaolues and henee no eigenved. 1
e
A |
481-je7-A|-
Eigen Values 4ltl=o t=t
R(Real numbey, hen A has o elgnvaue&
the field is
hence no eigenvect
In case field k= C Comblex field, then we have falbwing sesut
Theaen 4.t: ket A be 8quase
maix ovee the Comdlek field C.
Then A has a least ae eiqenvalue.

Theosen .3 Sukese i, Ya, -

Vare nonzero igenvectrs ofa matnix A

belomging to distiod eigenalues , da,*. n. Then
hen V1,Va;"V
Vi, Va, -V
Qve ineasly independen+,
(t-an. Then A
heosem 949. ut a(t)= [t-au)lt-4) -
l& Simila to the cliagmal mx D=daq/4,4,-,4n

T: V V
iagondli zatan oLinaas Cpaator
be sepresenfed a diagonal maik.D
is diaqorakizable f it can

Thus, T is diagoralizable iff 3 baais S =4, ,unt V

S.C.
TN)= R4
T)= R
te. O
TIL,)- kn Rn

D-deglska, ,Pn
Definhn kel TVV bek.T.. A calas is calld an eigen
Value T 3 h o exisk) a mzro vecor v s.t,

TCu= a0| igenvector

 AHgositm 4.7 (Diagmali zahon AHgonthm)
Thput An n- &quaie macix A .

Charadenishic polynomial 4/t/ of A,

Hnd the
td eigen Values
the A
Step2 Find the vook Att)
A.
too each eigen Value
Slep3 Repeat skp 3(o) 4nd 3(b)
A -AI Aubhading
@)Fo the mahix M=
daun the lingonal o A
the soludion space the harngenamus
() Hod a btat
system MX = 0

collection S= f v,v-., Vm?of al ciqenviech

Conaider the
Step obtuined in Step3
not diagmaltzable.
the A is
fmn,
A is diagomalzable.
f m = n, then
and P Lv va - -Vn

P A D= liagA, An)
comeaponding b the eigenvec V.
wheu ls the genvalue
at
(421
A31 t-3t-10 -a(t/
pltr-Al= |
Sp1e-l= t4 -2= (tH]/tu) -6 =
ttl
A(tl=o (e-3t-10)=lt-5)(tta)=o

Step 6-2,5
lt X= ( t
Skepa A=,
A-AT)X 0 >/6 °(6)
3
6x42 3+y det = t, y= -3t
3x
3x n 3+ y0=o
-Cb-3t, t#o
t
E Spunh-3) 1,-3
 M= A-5I= | 2

- 6J

MX#0 -+ 0 -2y=0
3t-6 l t y=t, 2= 2t
= (2t, t), fer t=|

Shep
iagonaizahon ofA PoL - L
p -21-f% -
- 2 lo
PAP - 4 2T 6
-/3 %
h -20
O

P'AP dg-2,5)
Is B diagonalizcble ?
Example 7
(t-5)(t-3)+1-t-8t+E
Cusateiche olynemial: ItI-s|
=
|t-5
t
Alt)= 0
&igo-alues
8tt =0 (t-4) o t4,4

EiganectsS M A-I
M-
 MX C
x- O 7-Y-0
X- o
ut y-t
-t
eigen vectesis = (t, t)=t0,)

So eign vectss oB vecks (1)

s mthpls e only and they
ahe not L.I.
B 16 not dugmaizable,
Example 9.8 Ie A diagonalizable ove R T
A3 -3 Is A diagonalizable oves C P

) Atl1tz-A= B*6 - 2 t+3||

=(t-4)+lo =o
t+1=0 t=tt|

Asvoot/eiperrs A hss no eiqenalues od no eigenveterk

i ConpleX numbers Oves the field R.
Ais not diagnalizable,

4tl= t4 = o t t E C.
dnd herce A
So A has two dishindt eigenvekks and -
hab
hab two independent eqenvectsk.
-5
Ma A-I =| 3-
2
3-
MX =o M= 0
MX= 0
) -5y=0
2 t3)y=0 So infinitely many &oluhions

lt t
Egen vects is -(C2, 1
 (26)
-5
M A-AI =|2+ MX =0
2 -3t TMI=O

(3ti) - 5y 0
t y t
2 +(-3+")y=o

IPl= +3 3-
P- = | Lus -

IP=i+3-3=t 2

- Pl-
L+3

AP 3

3-+3-c2
--3i+i*
-
2

3t-3t
-3¢t1+it3

d u gO ding lhsa
O -

F'AP diaa (A1,A2)