0% found this document useful (0 votes)
36 views

MTH308 Lecture 11

1) Crout's algorithm can be used to find an LU decomposition of a tridiagonal matrix A, where L is a lower triangular matrix and U is an upper triangular matrix such that A = LU. 2) Gaussian elimination with partial pivoting can be modified to solve systems of equations when the pivot element is zero. This is done by swapping the row with the smallest non-zero element in the pivot column with the pivot row. 3) Both Crout's algorithm and Gaussian elimination with partial pivoting are methods for decomposing or solving systems that involve tridiagonal matrices. Crout's algorithm finds the LU decomposition directly while Gaussian elimination solves systems of equations involving a tridiagonal matrix.

Uploaded by

ritvik goyal
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
36 views

MTH308 Lecture 11

1) Crout's algorithm can be used to find an LU decomposition of a tridiagonal matrix A, where L is a lower triangular matrix and U is an upper triangular matrix such that A = LU. 2) Gaussian elimination with partial pivoting can be modified to solve systems of equations when the pivot element is zero. This is done by swapping the row with the smallest non-zero element in the pivot column with the pivot row. 3) Both Crout's algorithm and Gaussian elimination with partial pivoting are methods for decomposing or solving systems that involve tridiagonal matrices. Crout's algorithm finds the LU decomposition directly while Gaussian elimination solves systems of equations involving a tridiagonal matrix.

Uploaded by

ritvik goyal
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 16

JLectare-l [P-41

Defntien (Symm etric matrix"


said to be
A matri A E Ma (R) is

etric matix A=AT (Trnapou A)


De ( Positve
A t M. (R) is Aoid to
At
A szmmetric matriw
be positive sumi deindte
)

Dei
pe (postuz defte) to
A szmmewic matri A ¬ Ma (R) is said

be positve defiwite if

(Sulmatric' 4Ss n
ut
Let KeN, 1sk< m and Cacdinality of «)
aith 1«= K (

A= (aj) ¬ Mnn (R). kxl pith indices


A subnati e o A oe ordorc
is a matx. LSe) E Me )
A
P-2 Sch thot
AL8,P] =
ts
|=p|=k,then

is min o.

winqmi)
The the matf

Called a
allad
Gnd deotminant of A]
nin or.
prinüpal

principal sabmarix
leding and detemiant AK
Called the ladigprincipal
ninor.
|Lectwe
2
5
7
A=o 3
35

AL) 3

A[]=

qrhinal abmatie

2
a of rdee 2.

2 3
8
G 3 5 fardue
3
3)
(pustpoua) pron
elenaats.
digo
oma
L=lss0e with
ottriangular
origatm. chlaty pes . has A
A=L A
ife deFuxte. postive A
is R)¬M netrix
ThesTemli
='-
L=
A A- LLT A=
A=
nm-gt ia A
is AThon
.positive- A
depnite. semiostve-
Fwethermate
if necescaray is Athen
Gndmmetie g
n chalaky A if that note We
ahasa
A=LL
the Aueh (R) LE matrin
Mn Anlarc triang
eists thera
MnE A matr A
have
a Beid
to is (R)
factorigtal)
Defintiam](chalaky
4
Lectre
Mn (R) be a ogmmetrie matri.
em Lot Ae
squivalent

() A ic positive dufirite
te matt AA
minors of
i) AN he prinipal
poseive matrik A a8ce
eigen-valnes f he
(3) l the
posttve

Find ch lesky deconfostom


Agorrt to
T4 T
A=
dat AEa,)- 4 ((2-4)- (243)
+H2- 4)
Note A= AT. = 32 -5 -6=21)0.

posstue dente.
’ Ais
4 T
2
4
2

( olumnasise o also Can be done)


Conaing toke y=-2.
(one Conld also
li= 4 ’ y=2 and pocoed)

+ =4 +-4’ t y = 7
One conld dlso
take

-,+J 2 2=24%-%
2Sig

3
t
27
20

Go-5-272% 20
20
- v e Can
( also be
Loci1 P-7
2

Goraal tgortha
1sisj Sn
A = LT
1$3 <i sn

(-Lij-iu)

k=

For i=1, and j=I


(
j- 2, -- n

.an.
tor i= 2, 3 - n-l

tor j=i, aii=) iik


) L+ Lii

for j= itl, , n

for i=n, j=n


Lote
Finally
i=j=|
For i=l, j=2, ,n

for
Lii-(O-ik)
1= it,. ,n

n-l

for i=j-n,

Tridigoral
A matrix p the form

A=
Ontn

|i-sl72.
We nsd sho a LU dulomoostan(Croutre)
assume trat
of atidagal matix (we
LU decomfoshon is qossbe
non-ero enties
A ha ot mast n+2(n-1) =3n-2.
One that the Cxouts deomposthi n
an shi
the frm.

A= LU whee
jo,ij
L L
L=

and i<j

A =LU
|Lotwne-1

(i-)A1

ai,i- = Lii-l or i= 2,3,,n

+ ii

for 1=2,3 nt

2 -1 0
0
u, oo
2
L 0

Li

Lg2 Lstss L s4

ComplaBe it
P-12
We nte that in Gaussan eliminti on
G ate K-yth
tsansfomd
(for 1sK Sn-t) the
akes the brm
matiz
a C.
(2)
(2)
()

a) a)

Onn
a e alleol
l=, , K-t. elemauts
ohe a o pivot

then
a=0 )

Can not to define the maltiplie


eimiate Rties in the Kth
to aoso the
ak)he
(K) wal .

Gawsan eimiatmn s soppad.


We neoal to
modify he method
Kth Column at A)
to sa, the
he ided is nth rod for
om the
hot non-gero
the Fist Smalest

(pith then

do the ntuabarg
Contnel.
and the ro s

This mehad is Cake pivoting


erro,it is Sometirme
reduco, ound op
lo irtarchanges
to porosm not zevo.
pivot elemnt- orte
if the magnitode conhors
moll n
) a is
the the magnitae
to )

the multiplive
wil be vey
Comona to otkre
elomants.
Hene
Corporated in the
is n
the tems
a-O.
P-14
baskd bsttutsn
when
(e)
(4() j=t
vy
awith Amallee Vale of
A
ortn the
to the numeatr, any
becase
Creased
numerator mioht be tn

Considee

Gawson emiraksn

ha 5-digit unling afpeniai n


2-Y% = 2-1o° =999998
e=10 Then, (-Y =l-|05-94999
and y=0.
Bwt and
Commits ).
Leet 1P5
Pocdkal pivoing
pivoting is paefosmad
To avoid this oro blem,
. a ) with

same ColumY hat


Seleet an elenont in -the
and has the
belo
absolute Valne.
KsPSn
Somaest
rallest itegsr
Lot p be the
Bh -that
ma a(9
Ksisn

r intu chargs)
alle portal pivotgbut mgh
This
(For beter aceroy
time regird)
Conporate he interehange
Can in
coluwns
bot ToO8 anel
choos mallest
rmall est mtagoe
)
(H) ma
K<isn
ksjsn
Qnol

(Gluamn ataccn

(2
(y
-
a

( (xy
(
6

Cameolemonts Comyarealeratr
here for hoe for

pioaing

You might also like