LinearAlgebraSummaryofResults

Compareresultswithsyllabusoutline(numbersrefertosyllabusTopic1)

1.Matricesandoperationswiththem,includinguseofGDC.

Payparticularattentionandthinkthroughthedifferentwaysofrepresenting
matrixmultiplication(seeHowardAnton,p.2934).Theywillbehelpfulfor
theoreticalinterpretationofmatrices,systemsoflinearequationsandvectorspaces.

Bycolumns,seeformula(6)onp.31.Thisisthemosthelpfulandimportantone
inthetopic!Itwillbeused(amongotheruses)alsointheEigenvaluespart.

MatrixproductAXasalinearcombinationofthecolumnsofmatrixA,(seebelow
(10)andExample8fromp.32.)

 Matricesasrepresentationoflineartransformations(seeparagraph1.8,p.76)

2.

Matricesinsystemsoflinearequations
Matrixofcoefficients,augmentedmatrixofthesystem.
MethodofGaussianelimination,rowechelonform,reducedrowechelonform
useofGDCtofindthem.Useofelementarymatrices,includingwithGDC.
Numberofsolutionsfromrowechelonformsolutionofsystemswithinfinitely
manysolutions.
Homogeneoussystem Ax = 0 andthesetofitssolutions(thenullspaceorthe
kernelspaceofmatrixA).LetAbeamatrixofdimension m n .
N (A) = N ull(A) = K er(A) = {x Rn|Ax = 0} isasubspaceinthedomain
ofthelineartransformationrepresentedbymatrixA.
dimN(A) = nullity(A) = n rank(A) representstheranknullitytheorem
(box1.6ofsyllabus).Itisalsoequaltothenumberoffreevariablesin
thesolutionofthehomogeneoussystem.
ThenullspaceN(A)andtherowspace R(A) = Span(a1, a2, ..., am) are
orthogonalcomplements.
Solutionofthesystem Ax = b .
Thesystemisconsistent(hassolution)if,andonlyif,
b C (A) = Range(A) = Span(a1, a2, ..., an) = {x|x = m1a1 + m2a2 + ... + mnan}
i.e.bisalinearcombinationofthecolumnvectorsofA.
Thesolutionofthesystem Ax = b canberepresentedas
x = "particularsolution" + N (A) (box1.7syllabus).

Example1:

3. Someimportantspecialcases
ConsiderasquaresystemoflinearequationsAX=B,whereAisa n n matrix.Thenthe
followingstatementsareequivalent(partofthemTheorem1.6.4,p.63)
a) Aisinvertible(hasinverse,isnotsingular,)
b) detA =/ 0
c) ThehomogeneoussystemAx=0hasonlytrivialsolution.
d) ThereducedrowechelonformofAistheidentitymatrixIn.
e) Aisexpressibleasaproductofelementarymatrices.
f) ThesystemAx=bisconsistent(hassolution)forevery n 1 matrixb(righthand
sideofthesystem).
g) ThesystemAx=bhasexactlyonesolutionforevery n 1 matrixb(righthand
sideofthesystem).

h) ThelineartransformationTdefinedbythematrixAisbijective(onetoone).
i) TherangeofTisRn.Thekernel(orthenullspace,thesetofallvectorsmapped
tothezerovector)consistsonlyofthe0vector,i.e.is{0}.
j) Thesetofthecolumnvectors(andtherowvectors)ofthematrixAislinearly
independent.
k) Thesetofthecolumnvectors(andtherowvectors)ofthematrixAisabasisin
Rn.

ConsidernowasquarematrixAwhichisnotinvertible(discussthepossibilitiesand
formulateasimilarsequenceofequivalentstatementsasabove)
a) Aisnotinvertible(doesnothaveinverse,issingular,)
b) detA = 0
c) ThehomogeneoussystemAx=0hasnontrivialsolutions,i.e.hasinfinitelymany
solutions.
d) ThesystemAx=bisconsistent(hassolution)onlywhentherighthandsidevector( n 1
matrixb.)isfromtherangeofthetransformationdefinedbyA,i.e.bisfromC(A)the
columnspaceofthematrixA.
e) ThereducedrowechelonformofAwillhaveatleastonerowofzeroesi.eitwillhave
theform
(withpossiblyrearrangementofthecolumns).
ThentherankofthematrixAisr<n.
i. Thismeansthatinthesystemofncolumnvectors(rows)ofthematrixA,
{c1, c2, ....cn} thereareonlyrlinearlyindependentvectors(theycanbe
identifiedfromthepivotalcolumns).Thoselinearlyindependentvectorsforma
basisforthecolumnspace,C(A),ofthematrix.
ii. Asimilarresult(thereareonlyrindependentrowsformingabasesinR(A)the
rowspaceofA.

Example2:
Considerthesystemoflinearequations:

Theaugmentedmatrixis

anditsreducedrowechelonformfromtheGDCis
.
Theinformationthatwecanreadfromtherrefis:
Column3ofthematrixisalinearcombinationofcolumns1and2,more
precisely(checkit!) c3 = .6c1 .8c2 .Inotherwordscolumns1and2forma
basisforthecolumnspaceC(A)andthethirdcolumnhascoordinates ( .6, .8)T
inthisbases.
Thesystemofthelinearequationsisconsistent(hassolution),asitsrighthand
sidethevector b = ( 1, 4, 2)T isinthecolumnspaceC(A).Indeedtherreftells
usthat(checkit!) b = 1.4c1 + 1.2c2 .
Thegeneralsolutionofthesystemis (x, y, z)T = (1.4, 1.2, 0)T + t(.6, .8, 1)T .Itis
thesumofthesolutionofthehomogeneoussystem(orthenullspaceofthe
matrixA) N (A) = {t(.6, .8, 1)T , t R} ,andtheparticularsolution(any
particularsolution),inourexample (1.4, 1.2, 0)T ofthesystem Ax = b .Thisis
ageneralresult(theorem)trueforallsystemsoflinearequations.

4.

Matricesasrepresentationsoflineartransformations

Definitionofvectorspace.SpecificexamplesRn,Cn,P2,M22,etc.
Themostimportantis Rn .
Wecanvisualize,orinterpretgeometricallyresultsin R2 thecoordinateplane or
R3 the3 dimensionalspace .
Whatarethepossiblesubspacesin R2 ?
Whatarethepossiblesubspacesin R3 ?

Lineardependenceandbasis.(Seeparagraph4.3and4.4inbook).

Example3:
Tocheckthatasystemofvectorsisindependent
1.Determinewhetherthevectorsarelinearlydependentin R3 iftheyarenotfindanexpression
forthelinearcombinationequaltozero.
a) ( 3, 0, 4)T , (5, 1, 2)T , (1, 1, 3)T

MakethematrixAconsistingofthegivenvectorsasrespectivecolumns,i.e.


Thequestionaboutthelineardependenceofasystemofvectorsisequivalentto:
findingoutwhetherthehomogeneoussystem Ax = 0 hasnontrivialsolutions.Anon
trivialsolutionwillprovidethelinearcombinationofvectorsequaltozeroandwillmake
thevectorslinearlynonindependent.
Anotherpossibleformulationis detA =/ 0 ?Ifyes,thenthesystemofvectorsis
linearlyindependent(andformsbasesinthecorresponding(sub)space.
Inthecaseofarectangularmatrixthequestioncanberewordedaswhethertherankof
thematrixAisfull(equaltotheminimumdimensionofthematrix).

Inalldifferentrewordingsofthequestion,asusual,ourmaintoolistherrefofA.Inthiscase
rref(A) = I ,(i.e. detA =/ 0 ,or rank(A) = 3 ),thereforethesystemofvectorsisindependentand
formsabasisin R3 .
2. Considernowthesystemofvectorsfrom R4 givenas
{(2, 1, 3, 0)T , ( 1, 1, 0, 2)T , (7, 5, 6, 6)T } .Thecorrespondingmatrixof

columnvectorsis

anditsreducedrowechelonformis

.
Fromtherrefwededucethat:
TherankofAis2,thisissmallerthan3(thenumberofvectorsinthesystem),therefore
thesystemofvectorsislinearlydependent.Indeedtherreftellsusthat c3 = 2c1 3c2 .
Thusthereareonly2independentvectorsinthesystem {c1, c2} andtheyformabasis
fora2dimensionalsubspace,Sin R4 (ortheyspana2dimensionalsubspaceS)namely
S = {m(2, 1, 3, 0)T + n( 1, 1, 0, 2)T m, n R}

b) Changeofbases(notincludedinsyllabusexplicitly,couldbeasked?)
Example4.
Giventhevector v = (6, 1, 3)T inthestandardbasisof R3 {e1, e2, e3} ,finditscoordinates
inthenewbases {b1, b2, b3} ,where
b1 = e1 + e2 + 2e3
b2 = 2e1 e2
b3 = e1 + e2 + e3 .

Thequestioncanberewordedasoneaboutasystemoflinearequations,(asalmost
always!)i.e.wearesearchingforscalars m, n, t suchthat v = mb1 + nb2 + tb3 = B(m, n, t)T ,where

thematrix
iscalledthetransitionmatrix B = T be frombasis {b1, b2, b3} tothe
standardbasis {e1, e2, e3} .Inthetransitionmatrixeachcolumnrepresentsthecoordinatesof
correspondingvectoroftheoldbasisthroughthenewone.Itiseasytoseethatthesolutionof
thesystemis (m, n, t)T = B1v = (1, 3, 1)T ,sothecoordinatesofvinthenewbasisare (1, 3, 1)T
(Checkit!).Notethatthematrix B1 = T eb .

Lineartransformationsandmatrices
Definitionandpropertiesoflineartransformation(operator)

Tofindthematrixrepresentingalineartransformation
Example:Considerthetransformation M y=x Reflectionintheline y = x (thisisthe
geometricdescriptionofthetransformation)
TofindthematrixArepresentingthistransformation:
Findtheimagesofthecoordinatevectorsunderthetransformation
and
.
ThematrixofthetransformationAisformedbytheimagevectorsascorrespondingcolumn
vectors,i.e. A = (e1|e2) .
So

andthetransformationisrepresentedbythematrixasfollows

(thisisthematrixrepresentationofthetransformation).

Whatcanyoureadfromthematrixofatransformation?

 Theimagesofthecoordinatevectors.
Thescalefactorbywhichtheareaoftheoriginalistransformed.
Areaimage = detA Areaoriginal
Compositionoftransformations(transformationaftertransformation)
Likeinthecaseoffunctionsweconsidertransformationsthataretheresultof
othertransformationsperformedoneaftertheother.Letthetransformation T A
representedbythematrixA,beperformedafterthetransformation T B
representedbythematrixB.Thenthecompositetransformation T B followedby
T A or T A after T B isdenotedby T A T B andisrepresentedbythematrix
AB .Notetheorder!(fromrighttoleft)

Famoustransformationsin R2 wearesupposedtorecognizeandknowthe
matricesofthefamoustransformationsintheplane

Geometry

Determinant

Rotationbyangle
clockwise

detR = 1

Matrix

Reflectioninline
y = tan.x

detM = 1

Projectiononline
y = tan.x

detP = 0

Projectiononline
y = mx

detP m = 0


Enlargement
w.r.t.O(0,0),
scalefactork

detEk = k2

 KernelandRangeofLinearTransformations

Example:
Findthekernelofthetransformation T = P m projectionontotheline y = mx .
Geometricallythekernelconsistsofallvectorsfrom R2 whoseimageisthezerovector,orthe
origin O(0, 0) .Thusthekernelconsistsofallvectorsonthelinepassingthroughtheoriginand
perpendiculartothelineontowhichweproject.
K er(P m) = {(x, y)T |y = ( 1/m)x}
Therangeofthetransformationcoincideswiththecolumnspaceofthematrixandconsistsof
thevectorsonthelineontowhichweproject.
Range(P m) = C (P m) = {(x, y)T |y = mx}

EigenvaluesandEigenvectors

EigenvalueisaGermanword,notevennearlypronouncedasifitwasanEnglishone.Itstands
forownvalue.Eigenvectorsformamatrix,whichdiagonalizesthecoefficientmatrix.

Theeigenvalueproblem:
Foranysquare n n matrixA,findthevaluesofthescalar suchthatthesystemofequations
Ax = x
hasnonzerosolutions.
Thesevaluesof arecalledeigenvaluesofthematrixA.
Thesolutionsofthesystemofequations(infinitelymany!WHY?)corresponding
tothevaluesof ,arecalledeigenvectorsofthematrixA.

Thesystem Ax = x canberewrittenas Ax x = (A I)x = 0 .Thereforethesystemisa

 ThisiscalledthecharacteristicequationofmatrixA.Itisapolynomialequationof
degreen.Wearerequiredtosolveonlycaseswhen n = 2 withreal,distinctroots.

Example
Giventhematrix
.
Thecharacteristicpolynomial P () is

Thecharacteristicequationis P () = 0 anditssolutions,theeigenvaluesofthematrix,are
1 = 5or2 = 2 .

Propertiesofcharacteristicequationandeigenvaluesproveandthinkthrough!
1 + 2 = a11 + a22 = T race(A) .Inotherwords,thesumofeigenvaluesequalsthesumof
thediagonalelementsofthematrix(thetraceofthematrix).Trueforalln,weneedto
knowandunderstanditforn=2.
12 = det(A)
ThematrixAsatisfiesitscharacteristicequation,i.e. P (A) = A2 3A 10I = 0

Thecorrespondingsolutionsare: y = 3/4x, when = 5 and y = x, when = 2 .Sothe

ThematrixPcontainingtheeigenvectorswilldiagonalizethematrixA.Finishtheexample!
..

Somesynonymsglossary

Topic(context)

Systemsoflinear
equations

Matrices

Linear
transformations

Solution(s)of Ax = 0 Nullspace N (A)

Kernelofthelinear
transformation
representedbythe
matrixA,Ker(A)

Setofvectorsbfor
which Ax = b has
solution(is
consistent)

Rangeofthelinear
transformation
representedbymatrix
A

Columnspaceof
matrixA, C (A)

Linearindependence SystemAx=0has
ofasystemofvectors onlytrivialsolution.

MatrixAofthe
vectorsascolumns
hasfullrank.