LINEAR ALGEBRA A High Quality Study Material for Higher Level Exam for V.G. L P.G. Students ( adis's Pliner Juste —” | Su es} | x D tps sswdony , | REGENERATING LOGICS \ An ee 9001 : 2008 Certified institute ay First Edition: Aug. 2018 Copyright © Dips Academy © Allrights reserved by Dips Academy. No part ofthis book may be reproduced or distributed in any formor any means, electronic, mechanical photocopying or otherwise without the prior permission of the ‘Dubey’s Information Pool & Solutions Pvt. Lig Dips Academy has taken due care in collecting the questions and providing the solutions, before publishing book. Inspite ofthis, ifany inaccuracy or printing eror occurs then, neither Dips Academy norits author shallbe responsible, However Dips Academy willbe gratefullfyou could point out any such error. Your suggestions willbe appreciated. L Scanned with CamScannerChapter 1: Linear Transformation and Its Properties, 1.1 Definitions 1.2 Range and Null Space of Linear Transformation ... 1.3. Sylvester's LaW ene 1.4 Types of Linear Transformation .. 15 Properties of Linear Transformations 1.6 Some Important Linear Operators. — 1.7 Set of all linear transformations from V" to" 1.8 Propositions .. 1.9 Isomorphism of Vector Spaces 1.10 Matrix Representation of Linear Transform: Chapter 2: Matrices and Their Properties 24 Definitions 2.2 Special Matrices .. 23° Rank and Nullity of Matrices 24 Properties of Matrices .... 2.5 Properties of Determinant 2.6 Properties of Inverse of a Matrix 2.7 Properties of Rank of a Matrix .. 2.8 Properties of Trace of a Matri 2.9 Properties Important Matrices/ Linear Operators... System of Linear Equations.......... Scanned with CamScannerChapter 3: Eigen Values and Eigen Vectors Definition: Characteristic Poly nomial .. Eigen Values and Figen Vectors Similarity of Matrices and Dingonalization oo. Spectrum of Figen values Dingonalizability of Important Linear Operators / Matrices Canonical Forms Jordan Canonical Form Scanned with CamScannerERC TELL Nee cue ce CHAPTER 1 LINEAR TRANSFORMATION AND ITS PROPERTIES 1.1 Definitions Let P(E)and P(A) be vector spaces (over Field F), We call a function T:¥ ¥V" alinear transformation from V(F) to V'(F) if, for all sy eV and ce F, wehave 1. T(xty)=7(x)+7(y)and 2 T(cx)=cT(x) Identity transformation /(x) = x is linear transformation from onto 7 Zero transformation defined by O(x)=0 is called trivial transformation. Example: RY ={(a,,a3,-dy)|qj ER} be a vector space over field R then the in 2resdn) | following are linear transformations: 1. T:R" 9R" such that 1 (4,4... tn) = (aysytost) 2. TiRN RM such that 74}, ,.44q)= (24 +43942 +43 rnd +) 3. T:R" RM such that 7 (4,425.04) = (44s 44250y-1,0) 4. TR" 9R" such that 70,495.) = (O,45-94y-4) Example: R'™" ={4=[ay]| a,j €R }be a vector space over field R then 1. 7:R™"—R™" such that T(A)=A": where A? denotes the transpose of A Atal 2 7:R™" +R" such that T(A) where A? denotes the transpose of 4 3, T:R™" R such that T(A)=T+(4) where Tr(A) denotes the Trace of A 4. T:R™" —R such that T(4)= ay jaliet ae RE =a Scanned with CamScannerPare Matrices and Example: R{x)={p(x)|p(2) is a polynomial over field R} then the Put Your Own Notes following are linear transformations: T:R[x]—> Rix] such that 7(p(x)) = p'(x): where p'(x) denotes the derivative of p(x) 2. T:R[x] > R[x] such that T(p(x)) = p"Ca): where p"(x) denotes the second derivative of p(x) 3. T:R[x]>R{x]such that 7(p(2))=[p pode 4. T:R[x]>R such that T(p(x)) = pO) 5. T: R[x] B such that T(p(x))= p'(0) Examples: 1. Ty :R? > R? such that 7p :(a,,a2) is the vector obtained by rotating (a),a2) counterclockwise by @ if (a),42)#(0,0), 7o(0.0)=(0.0) ‘Then Ty :R? — R7is a linear transformation called the rotation by 8 (a,c0s0—a, sin0a, sin +a, c0s0) T,(a.4,)= 2. 7:R? + R? such that T(a),a2) =(aj,~az). T is called the reflection about the x-axis 3. T:R? > R? such that 7(a,a2)=(4),0). Tis called the projection on x-axis along y~axis Example: T:R"™" > RI" T(X)=AX where dy, is fixed is a linear transformation, Note: 1, 3 an additive function which is not linear. C(C)+C(C) defined by 7 is not linear Example: Since T(iz) + iT(2) 2. T:M,(C)>M,(R) is mot a linear transformation as transformations are defined over same field. 3. T2M,,,(F)M,..(F) Let B bea fixed mn matrix. Then T(4)=4B-BAVAEM..(F) is always 4 linear | transformation | 4. Let 7,Q:V(F)>V"(F) be two linear transformations, Let a, BeF then T(x)=a7;(x)+B7,(x) ¥xeV is always a linear transformation from 7" ‘ips Hess Din Sara nes Khas Near LLT™ Nem De 110016 PR: (11) 3087797, Clr SITIONS Tih lnleadiomendems em: Web: wwe dlserndrer ce Scanned with CamScannerProposition: Let T:V(F)—>V"(F)bea linear transformation. Then (@) 7(0)=0: where Oon the LHS. is zero vector of V and Oon the RHS. @) T(> © T(x-y)=7(x)-T(y) VanyeV the zero vector of V7" T(x) VreV @ Paix Har +--Gnty) OT (a) + O27 (x2) +--G 7 (an) va €V,a;€F ,i=lt0n Examples: G) T:R,(R)>F,(R) defined as T(p(x))=fp(0)al is not a tinea 0 transformation since 7(x") =~ nat e(R) (i) T:R™™" (R) > R(R) defined as T(A)=det A(or [4|) is not a linear transformation, Since v 4,BeR™" (R) T(4+B)=T(A)+T7(B) does not hold. 12 00 ua de! 2] 52 walt alot |] 1.2 Range and Null Space of Linear Transformation Range space of a linear transformation: Let (F)>V'(F)be a linear ‘transformation then the Range of T written as R(7) is the set of all vectors Bin V'such that B=T(x) for somexeV. ieR(T)={T(x)eV':xeV} Range of a linear transformation 7 :V — V"' is subspace of ¥' Rank of A Linear Transformation 7: The dimension of range space R(T) of a linear transformation is. called Rank of a lincar transformation T. denoted by p(T). Rank of zero transformation is 0 Null space of a linear transformation | Let 7:V(F)->V'(F)be a linear transformation then the null space of | written as N(7)is the set of the vectors xis ¥ such that T(x)=0 (zero | vector of ¥'). Thus N(T)={xeV|7(x)=0} Nal space of T (Kemel of T denoted by ker()) is a subspace of ¥ ‘Dips Howse, 2, Sara tne Khas, Near LT, New Dei-110016 Phi (RTE. Cae a0 Ermalafoioacadem com, Webule! ww diveedemn co oe Scanned with CamScannerNRE Re eo eu Nullity of a linear transformation T > ‘The dimension of null space 'N(Z) or ker(T) of a linear transformation is called Nullity of @ linear Put Your Own Notes transformation T denoted by 1(7) Domain eee Co-domain ‘Null Space Range Space Sylvester’s Law Let V and¥” be vector spaces over the field F and let T be a linear transformation from V into 7” Suppose that V is finite dimensional then p(T) +n(7)=dim(V) This is also known as Rank-Nullity Theorem Result: Let 7” and V" be vector spaces, and let T:V(F)—>V"(F) be linear. If B={¥,.¥2s=s%} iS a basis for V, then R(T) =span(P(B)) =span({7(04).7(%2)o-»7 (%)}) Examples: 1. Define 7:IR?(R)-> R?(R)as T(a,b)=(a+b,a-b,b) We can easily verify that Tis a linear transformation, Now, {(1,0),(0,1)} isa basis of R?(IR) 1,1) The vectors T(1,0),7(0,1) span R(T) thus Range (7) is a subspace of R°(R) spanned by the vector {(1.,0), (1-10)} ‘There vectors are linearly independent. -. they form a basis of R(7)) => Rank(T)=dimR(T)=2 By rank Nullity theorem, —_p(7)+n(7)=dim(R?(R)) =>n(T)=2-2=0 null space of T must be the zero subspace of R? (R) OTHERWISE: Let (x,y) €N(T) > 7(x,»)=(0,0,0) =>(x+y.x-y.y)=(0,0,0) = x=0,y=0 (0,0) is the only vector of R? (IR) which belongs to N(T) Scanned with CamScannerT:R" +R" such that T(a,42,-++s4,)= (0,0, Then p(T)=n-1 .n (7) 3. TR? >? such that T(a,a2) =(a),-ay) Then n(T)=0, p(T) 4. T:B,(R)>P,(R) such that 7(p(x)) = p'(x) Then n(7)=1, p(T)= Result: Let dim(/))=n, T:V(F)—>U(F) be a linear transformation, such that the range and nullspace of T’are identical. Then 7 is always even . Example: T :R?(R)— IR? (R) be defined by T(a,b) =(b,0) V a,beR Then R(T)={(x,0)|xeR} and N(T)={(x,0):xeR} Types of Linear Transformation Linear Operator: Let V(F) be a vector space then any linear transformation from Y into V is called linear operator on V. Let ¥ and V” be vector spaces over the field F, let T:V—>V" be a linear transformation Tis said to be one-one (injective)if ker(T) = {0} Tis said to be onto (Surjective) if R(T) Tis said to be Bijective (Invertible) if T is both one-one and onto If T is one-one and onto, then the inverse function 7 is a linear transformation from "into V ‘The necessary _and_sufficient_condition for a linear transformation T:V—>V" to be invertible is that there exists a linear transformation S:V'¥ such that ST=1=TS. Singular: A lincar transformation is called singular if there exists a nonzero vector x such that 7(x)=0 ie, a linear transformation is singular if ker(7) #0 that is (7) 21 Non-Singular: A linear transformation is called non-singular if T(x)= 0=> x =0 ie, a linear transformation is non-singular if ker(T) = (0} Composition of Linear Transformation: Let T:/(F)>V"(F) and S:V'(F)>W(F) be linear transformation. Then, Sor:v >W defined by (SoT)(x) = S(7(x)) vreV is a linear transformation. igs Howse 28; da Sara Haut Khas, Near LT, New Dl "Ema inoedibacadery om Wl ‘Scanned with CamScannerWet etee EWU e or Tate Identical Linear Transformation: Let = {¥,¥2,.u¥,} be a basis of V If > STV V" are linear and S(v,)=T(y,) for i=1,2,.40, then S=T ie. if ‘wo linear transformations agree on basis then they are identical Negative of Linear transformation: Let V(F)andV'(F) be two vector spaces. Let T be a linear transformation from V into V'. Then -T is defined by (-T)x=-[7(x)]¥ xeV is a linear transformation from V into a Example: Let V=P,(R) and W=M,,,(R) then 7:¥—>W such that {° (0) 0 p)-pQ), Consider standard basis p= {1,x,x7} then R(T)=span(7(B)) then rank of Tis number of linearly independent vectors in T(B). at 10 »_{1 0 we ror-() )7e-(5 ‘). Tet=[, 4) >, there ate only evo linearly independent vector in T(B)hence p(T)=2and. (7) T(r) ) is a linear transformation, implics T° is neither one-one nor onto. Proposition: Let V and V'be two finite dimensional vector spaces such that dim V = dim Vand 7:V — Va linear transformation. Then the following conditions are equivalent: 1. T isinvertible 2.1 isnon-singular 3. T isonto AL TE {vy Vayecssson Uy i8 a basis of V , then {7(y), T(v2)s--u T(v)} is a basis of V*. 5. Tis 1— lor injective 6. ker(7)={0}= VM) 7. T carries each linearly independent subset of Y onto a linearly independent subset of 7. Important Note : In the above theorem, we have proved that if Tis a linear transformation on a finite dimensional vector space V, then Tis one-one implies T must be onto. Also T is onto implies T must be one-one. However this theorem fails if Vis not finite dimensional. This is clear from the following example. Example : Let V(F) be the vector space of all polynomials in x with coefficients in F. Let D be the differentiation operator on V. The range of D is all of V and so D is onto. However D is not one-one because all constant polynomials are sent into 0 by D. ip Hows 26; Js Sara Hana Ka, Near LLT, New Deni 110016 Ph ‘Ermallafozdioncadem cou: Website wwe dinucadetom Scanned with CamScanner DRS, Ca BOTTING gaRe etsy oak 0 A eo (igovenr toon Cectied sti 1.5 Properties of Linear Transformations Let V(F) and V'(F) be vector spaces. Let T:V(F)->V'(F) be a linear transformation Let T:V(F)-»V(F) bea linear operator If is finite-dimensional, then Rank(7)-nullity(7)}=dim(/) implies Tis bijective. Let T:V(F)>V'(F)be a lincar transformation (@) If x,x2.4%4 are linearly independent vectors in Y then T(x,),T(2),--T (az) may or may not linearly independent in V" (b) If 7(x),2(%p)0T (xq) are linearly dependent in" then 21.X2.-%q May or may not linearly dependent veetors in V - Example: If T'is trivial transformation T(8)={0} B being basis is L.1, {0} is LD (© Hf xyxgs-5x4 are linearly dependent vectors in T(x),7 (2) (xy)are linearly dependent is V" (@) If T(%),7(x2),..,7Q4) are linearly independent in V" 21stgsetp ate linearly independent vectors in . Let T:V(F)>V'(F)be a one-one linear transformation (@) Aexysey are linearly independent vectors. in 7 iff T(x,).T(%3) TC) ate linearly independent vectors in V". ©) yxze0% are linearly dependent vectors in iff. T(x),7 (%p)so-sT(4y) are linearly dependent vectors in V*. |. Tis one-to-one if and only if T carries linearly independent subsets of V onto linearly independent subsets of V". B={v¥2,0%,} is a basis for V and T is one-to-one and onto then 1(B)={7(4),7 (%2) eT (vq)} is basis for V" » Let V(F) and 7(F) be finite-dimensional vector spaces and T:V(F)V'(F)be a linear transformation If dim(V) < dim(V) then T cannot be onto. Ifdim(V) > dim(V'), then T cannot be one-to-one. + A linear transformation T is defined on R'(R) by 7(a,b)=(aa+Pb,y2+86) where a,f,1,8 ate fixed elements of a ‘Then T is invertible if and only if apy +0, ‘Ho 3 Ha Sr Hausa New ELT. Rew Doar 008 PLIST Ewin slowendent om: Webute: wwe dnmtnaverge OE BOG Scanned with CamScannerMatrices and System of Equations Some Important Linear Operators Projection: Let y=? be a vector space over field R consider the function 7: R? 5 R? 1. Tis projection on x~ axis along y— axis then T(a,5)=(a,0) for all (a,b) €R? is a linear operator Tis projection on y—axis along x— axis then T(a,b) =(0,b) for all (a,b) € R? is linear operator Tis projection on y=mx axis along y=—1x axis then m rie =( 228 m(a+mb) ) for all (a,b) € R? is linear operator Note: In general projections on any line is not a linear transformation but projection on y=mx along y =——-.xis always a linear transformation, © If we define projection on y=mx+calong cj if cH 0it is not a linear transformation as 7'(0,0) + (0,0) Orthogonal Projection on Subspace: Let V(F)be a vector space over field F and let W be subspace of V(F), define T:V + V such that TOw+w)=w, where V=W OW", (Wis orthogonal complement subspace of W’ ) then Tis a linear operator V’, called orthogonal projection on subspace 17 Ker(T)=W' and. Range(T)=W 1(7)=dimWV ‘and p(T) =dimv If W #V then Tis neither one-one nor onto , i.c., Tis singular linear operator 1? =7 ,ie., T is idempotent linear operator Example: V=C™'(R), Let T:V-V, be orthogonal projection on w={deV|4" =A} then R(T): {set of symmetric matrices of ¥ } ker(7)=W'= {set of skew symmetric matrices of 7 } p(7)=dim(W)=30 n(7)=dimW'=dimV -p(T) =50-30=20 Question: Find orthogonal projection —_ of, W =span{(1,1,1),(0,1,1)}. on 7, Ces BBU0221386 Scanned with CamScannerSolution: x=(1,0,1)¢W dim(I7)=2, dim(R3(R))=3, Hence dim(IV" Let W'=span{(a,B,7)} ‘Then (a,f,y).(1,1,1)=Oand (a, B,7).(0,1,1)=0 => a =0, 8 | = W'=span{(0,1,-1)} Now x=w+w"' wherew=(1,4,¥4)eW, (0-4w)ew" Ta) =rOvsw) =w=(1 4.14) Reflection Let _V=R? be a vector space over field ‘consider the function T:R? > R? } 1. Tis reflection about x axis along y-axis then T(a,b)=(a,-b) for all (a,b) € R? is linear operator T is reflection about y-axis along x~ axis then T(a,b)=(~a,b) for all (a,b) € R? is linear operator 1 —x then m To) (2( 22a -a( meant) ) forall (a,b) ¢R?is linear tnt +m | T is reflection about line y=mx , along y= operator T is reflection about origin then T(a,6)=(-a,~b) for all (a,b) R? is linear Operator S If isReflection about y=mx+e along y=—*+0 (c0) m | This Tis a function not a linear transformation ** T(0,0)+(0,0) Reflection on Subspace along its Orthogonal Complement: Let V(F)be a vector space over field F and let W be define T:V — V such that T(w+w')=w—w', orthogonal complement subspace of 1¥’) then subspace of V(F), where V=WOW', (Wrig Tis a linear operator on ¥ , W along W" Ker(T)={0}and Range(T) = n(7)=Oand p(T) = dim and is called reflection about subspace T is one-one and onto, i., T'is Invertible linear operator 7? =1,ie,, Tis Involuntary linear operator E-mail: afeioucademycom: Websites ge eh: OM RT es rae COE Haag canned with CamScanner \ Dip own sa He Ks Near LOE RODMatrices and System of Equations Question: The reflection of (1,0,1) about Ww =span{(1,11),(0,11)} alon8pye your Own Notes ‘orthogonal complement subspace of 1 ———— Solution: x=(1,0.1)¢W W'=span{(0,1,-1)} xewsw'; Where w=(1¥4.¥)e > w=(0-% Kew" T(x)=T(wew) =w-w'=(110) Rotation: Let V=R? be a vector space over field R consider the function 7:R2-3R?, T is rotation in anti-clock wise by\an_ angle 9 then (a,b) =(acos@—bsin0,asin0+bc0s@) for all (a,b) eR? isa linear operator 1. Ker(T)={0}and Range(T)=V 2. n(7)=Oandp(T)=dimY 3, Tis one-one and onto, i., Tis Invertible linear operator 4. If Ty & Ty are rotations in ant-clock wise direction by an angle Gand g then Tj2Tpis rotation in anti-clockwise direction by an angle O+@ Also,7, » Tp is rotation in anti-clockwise direction by an angle @ + ;. Rotations commute ie. Ty °T%p=T, °T Tis rotation by an angle Othen 1" =Ty °Ty °° To = Tho T is rotation by an angle @then ty =I for some ne Niff Qis rational multiple of x ,. If Gis rational multiple of xand T is rotation by an angle @such that hen o=i() Let V= be a vector space over field R consider the function T:RoR, T is rotation in anti-clock wise by an angle 0 in xy~ plane then (a,b,c) =(acos0—bsin8,asin0-+bc0s0,c) for all (a,b,c) V such that Ty(X)=AX for all X =(xj,x}.0%,)" Ris linear transformation, a linear Let TF" — F™" such that T(X)=AX where A, is fixed transformation. Moreover , (p(T) =m.p(A) ; n(T)=m.n(A); Trace(T) =m. Trace( A) Ais singular iff T is so. Let V=R"™" be a vector space of real matrices over field R and let AeR™" define then Ty:V—>V such that 74(X)= AX ot T4(X)=%A forall X ¢R™" both are linear transformations Set of all lincar transformations from V to V" Let V and V" be vector spaces over F. then set of all linear transformations form V into V' is denoted by L(V,V"). In the case that /=V" , we write L(V) instead of L(V,V') =R[x], and for 21 define 7;(p(x))= p(x), where Example: Let p(x) is the j" derivative of p(x) then the set {7j,FayJq} is a linearly independent subset of L(V) for any positive integer n Corollary: Let V(F) and V'(F) be finite-dimensional vector spaces of dimensions n and m , respectively. Then L(V,V") is finite-dimensional vector space of dimension mn If |F| = p (finite), then total number of linear transformation is p”” Example: dim (L(R(R),R*(R)))=3-4=12 But |z(R?(R),R‘(R)}|=infinite ‘Theorem: Let V() bea finite dimensional vector space Let {04,04,,.0,} be a ordered bases for /. Let (FF) be any vector space. Let yi, Ya Yn €V (be any n vectors) ‘Then there exists a unique linear transformation 7:V(F)—>V'(F). such that T(a,)=y,,1s isn Importance: If |/(FY or “(Fis infinite, then 3 ite number of linear transformations from VV’, where V #{0},V" {0} Propositions 1. Let V be a finite-dimensional vector space, and let 7:V — V be linear (@) If rank(T)=rank(T?) then R(T)MN(T)={0} and V=R(T)®N(T) ‘ips Hwan, 38; a Sarl Hau Kha, Near LT New be 0046, Ps (Dj ROA, ema lfeaiacadem com, Webate: wwrasnsrsenn A AOS eae Put Your Own Notes&) V=R(T*)@N(T*) for some positive integer k 2 Let T bean invertible linear transformation fiom Vto W.Tha io _——— finite-dimensional if and only if 1 is finite-dimensional and dim(V) = dim) 3. Let T:V 91 be a linear transformation. then following holds. (@) N(T)=N(-7) c () 4. Let Tbe a linear operator on a finite-dimensional vector space’. Then following results hola. @ v(r)en(r)c sou (r*)on(t")c @ (ri () If rank (1) = rank (1) for some positive integer m , then rank(7") = rank(1* ) for any positive integer k>m © if rank (1) = rank (r™") for some positive integer m, then n(z") N(*)for any positive integer k>m S. If T and S are linear transformation (on the same vector space), then a necessary and sufficient condition that both T and S' be invertible is that both 7S and ST be invertible. If T and Sare linear transformations on a finite dimensional vector space V such that 7S=Ithen T and Sare invertible and7™ = S NoTE: Above can be false when V is not finite dimensional. Example 1: Let V be the vector space of real valued sequences. T,, T, : V(R)—> V(R) be two linear transformations defined by T,(4,,4,,0,...)=(0,4,,4,,4,...) (Right Shift Operator) =(a,,0;,4,..) (Left Shift Operator) Then ToT, =I , But neither 7, nor 7, is invertible Since (1,0,0,..,0)€R(,)= 7, is not onto (4,0,0,..,0)€N(T,)=> Ty is not one-one Example 2: Let V(R) be the vector space of all polynomials in x with elements in R . Consider the linear transformations D and T on V defined as follows. Dv@-Zse) vs e)er And TLF(x)]=[f f(x)de ve (x)ev Pee elaine Web re ssadoasans Scanned with CamScanner mereHere DT=, But D is not invertible because D is not one-one T:V (F)>W(F),S:#(F)>Z(F) be linear transformations then (@) If SoT is one-one then Tis one-one $ may or may not be Example: TR +R? defined by T(a)=(a,0) SR? OR defined by S(a,b)=a Then SoT(a)=a¥ acR =SeF is one to one but kerS={(0, b):beF} = S isnot 1-1 (b) IF T isnot one-one = SeT cannot be one-one (©) SeT isontothen $ is onto T may or may not be i. Example : as in part (1.) ST is onto ii, But T isnot, as R(T)={(4,0): aeR} +R? (@ IES isnot onto > SoT cannot be onto (© If S and Tare one-one and onto then ST is also bijective S:R" 9R".T:R" SR", men (@) if SoTis 1-l=>mn>m (© if ToSis bijective = S is one-one and T is onto => n § can never be onto and T can never be one-one => SeT is neither 1-1 nor onto. © if SoT bijective = T'S neither 1-1 nor onto, Example: if $:R' + R', T:R*—>R? then SoT cannot be one- ‘one and onto If TVy Vp 5S Vy > Vp then () rank (Sor) = rank (T)~dim(ker(S)tm(7)) (b) If Sis 1-1 rank (S eT) =rank(T) (©) rank(S)+rank(T)—m< rank (S°7)V" be linear. Then for each j,1 R? be the linear transformation defined by T(a,,a,)=(a, +3a,,0,2a,—4a,) T(1,0)=(1,0,2)=le, +0e, +2e, and T(0,1)=(3,0,-4)=3¢, +0¢, —4e, 13 Hence [7], =|0 0 2-4 2 Ifwe let y'={e,,¢.¢}. then [7]. A 1 Matrix representation of some important Linear Transformation Let T:R? > R? 1. Tbe the reflection about x—atis i.e, given by T(a,b) Scanned with CamScannerMatrices and System of Equations 7(0,1)=(0,-1)=0.(1,0)-1.(0,1) °. the matrix representation of 7 w.r.t. standard basis of R? is given by 10 [ Similarly matrix representation w.r.t. standard basis of R? -1 For reflection about the origin T(a,b)=(-a,-b) is given of ri 0 For reflection about the line y=x, 7(a,b)=(b,a)is given oft ci A rotation operator in R? that rotates a vector through angle @ in anticlockwise direction is given by Tq (a,b) =(acos0 —bsin#, asind +bcos6) ‘The standard matrix representations of rotation operation in R? is [nm] cos? sind #1" | sind cose |. The matrix representation of 7:R"R" defined by T (0.43 ...dy)=(054,--y-1) Wert the standard basis is, fe 1 o oo 00 10 Results: 1. Let V and W be finite-dimensional vector space with ordered bases and y , respectively, and let T,U:V +17 be linear transformations. (©) [et], =a{7V, for alt scalars a 2. Let T bea linear operator on an n—dimensional vector space and let B be an ordered basis for ¥. Then T is invertible iff [7], is an invertible matrix. Also if 7 is invertible, then [7~'], =[7] ive, that matrix of 7” relative to B is the inverse of the 2 mate relative to B min ip Wan Fin aa Hans hs New UT New Da LTOOG Pas GUD TOT "Ema lloaloacadem Zon: Webs: ww dine nage CoE OIG Scanned with CamScannerMatrix multiplication is compos Let 7: mn of linear transfor M(F)V'(F) and S:V'(F)-W() be linear transformation Let B, B', BY be ordered hasis of 1”, I A=[TB:B]. B=[5:9':f"] Then Sc7:(F)1"(F) “IW respectively If C=[SoT:B:B']. Then B.4=C Similarity of matrices : Let 4 and B be square matrices of order 1 over the field F . Then Bis said to be similar to if there exists an nxn invertible square matrix CC with elements in F such that B=C4C* Similarity of linear transformation. : Let 7, and 7, be linear transformations on a vector space V(F). Then 7; is said to be similar to 7; if there exists an invertible | linear transformation S$ on V such that 7; = $7.5" Results: 1. The relation of similarity is an equivalence relation in the set of all linear transformation on a vector space V(F) 2. If two linear transformations 7; and 7, on V(F) are similar, then (7,)° and (7,)° are also similar and if, 7, ,7,are invertible, then (71)",(7,)" are also similar. 3. Let T bea linear operator on an n—dimensional vector space V(F)and let Band ' be two ordered bases for . Then the matrix of relative to iis similar to the matrix of T relative to =[T:':B'] Hence these matrices will have same Importance (calculations can be simplified) Example: Let T:;(R)->P,(R)be the linear transformation defined by 7(r(x))=P'(*)- Find trace [7':: B] when P ={6,3+4x,2+24417 Solution: Instead of finding [1':f:f]we can easily find [7': f':f'] where B= {hax} 0 1 2) [7:B:B]=|0 0 2| Hencetrace [7: p: 000 Note : (T: B: B'|nced not be similar to [7: B': Sa RS a RT RT { carat ania as ee Scanned with CamScannerWerner nine kaa erent Examples: Consider a tinear T(xy)=(n.x). Let =f basis of R?(R). atroealf 4] Belrpeml[3? transformation 7:R? >R? given by (1,0),(0,1)} & Bz ={(1,2),(3,4)} be two 3 “1/201 de(4)=2, det(B)= 4 Hence they are not similar. 1. Let T bea linear operator ona F.D.V.S.V 1 of T relative to any ordered basis for TI (@) det(r) =der4) (F). Let A be the matrix hen, () Rank(T) = Rank(A) © Trace(T) = Trace(A) + Let T ands be linear operators on the finite dimensional vector space V(F), Then (@) det(7S)=(det7)(det s) () 7 is invertible iff det 7 +0 © IfTS=0,7 40,540 then det T =dets =0. (@) If {oy,.0,} and {B,,..B,} ate bases in the same finite dimensional vector space V(F), and if T is a linear transformation such that 7(a,) =p, i=Loom, then detT £0 Put Your Own Notes — rte a ae eT RNG " Esl nodoneaden cont Went wg IECHAPTER - 2 —<$<_<— MATRICES AND SYSTEM OF EQUATIONS 2.1 Definitions Let Aprun =[4j] be matrix with either real entries or complex entries Row Matrix: A matrix having a single row is called a row matrix. Also called row vector . Column Matrix: A matrix having a single column is called a column matrix. Also called column vector Square Matrix: An mxn matrix A is said to be a square matrix if m=n ic. number of rows = number of columns. The elements a, for which i= j constitute its principal di: Diagonal matrix: A square matrix is said to be a diagonal matrix if all the elements lying above and below the principal diagonal are equal to 0. Example: 300 A=|0 5 0 | also denoted by 4=diag(3,5,-1) 00-1 Equal Matrices: Two matrices are said to be equal if they have the same order and each element of one is equal to the corresponding element of the other Unit Matrix or Identity Matrix: A diagonal matrix of order n which has unity for all its elements, is called a unit matrix of order n and is denoted by J, Null matrix : The mxn matrix whose elements ate all zero is called the null matrix (or zero matrix) of the type mxn Triangular Matrix: A square matrix in which all the elements below the principal diagonal are zero is called Upper Triangular matrix and a square matrix in which all the elements above the princi are zero is called Lower Triangular matrix. Note: 1 diagonal (a) Diagonal matrix is both upper and lower triangular (b) A triangular matrix A = [ajo is called strictly triangular if ay for! rank(A) =3 Properties of Matrices 1. 4+B=B+4 [Commutative] 2. (4+B)4+C=4+(B+0) [Associativity] 3. A+O=A=O+A — [Existence of Identity] 4, -A+A4=O=A-A — [Existenceof Inverse] 5. A+B = A+C > B=C > B+A=C+A ‘Matrices and System of Equations Put Your Own Notes | Dips Toes 28 Oe m7, dean oo ast Ca 1] Scanned with CamScanner=> B=C —— [Concellationlaw] ‘The equation A+X=0 has a unique solution in the set of all A( mxn) matrices K(A+B)=kA+KB [Kk be any scalar] (p+9)4= pA+a4[p,q are two scalar} (pq)4= p(a4) 10. (-K)4=~(KA) = K(-A) MM. 14=4,(-1)A=-4 12, -(A+B)=-4-B 13. If the product AB exists, then it is not necessary that B4Product also exist. 14. A(BC)=(AB)C [Associativity] 15. A(B+C)= AB + ACIdistributive] 16, The equation 4B=0 does not necessary Imply that at least one of matrices A and Bmust be a zero matrix. 17. If AB=0, then it does not necessarily imply that BA =0 18. Al, =A=1,A 19. 42—B? =(4—B)(448) is true only if matrices 4 and B commute 20, The Product of two triangular matrices is itself triangular. 21, If'a diagonal matrix is commutative with every matrix of the same order then it is necessary a scalar matrix. 22, If AB=A and BA=B then A and B are idempotent 23. wr(kA)=k.r(A) 24, tr( A+ B)=tr(A)+ir(B) 25. tr(AB)=tr(BA) a6. (4?) = 27, (A+B) = AT BT 28, (4B) = 87a" A Bi) q Db)" 29. Let a-( Ady +BCz AB +B,Dy Cy +DiCy CyB + DD, ‘A BY (AT ct 30. = - cD, BT DI a Sara aus Kis, Near UT, New Deh-10016 Pas (02897327, Cale BOOTDLI Temallfoadlonaculen cou: Webste: ww dlosacedea. co 2=20-( Scanned with CamScannerMatrices and System of Equation A M. Let z-(é 2 Where A and D are square matrices, not Put Your Own Notes necessarily of the same order. (a) If Z is symmetric then A and_D are symmetric. | | (b) If Z is diagonal then 4 and D are (©) If Z is upper triangular then 4 and D are upper triangular Converse of all the above 3 statements need not be true. gonal (@) (4°) <4 ©) Cas B= a eB! () (ha) = ka (e) (48) = B°a® (hy 1r(4)® + 174 Gin general) 33. (4B) = 4"B" (if AB=BA) 3M. For real matrices 4,8 and ( @) 47420 if A=0 (b) aB=0 iff 47 45-0 (©) 4B=AC it 47 4B= Al AC 38. (aa)! = 478) intag = Ba 36.1f 4 and B be diagonal matrices then 48 is also diay. aaa 1B is also diagonal and that 37. For real matrices 4,8 and C 4 4=0 ifandonly if 4-0 AB =0 ifand only if 4148 <0; AB= AC ifandonly if 4148 = 4°4C Note Result docs not hold if matrices are not real for exam for cxample take Scanned with CamScannerz,_( 4+ BiC2 Br + BD M2" CA, + DiC) CyB, + DDy, Properties of Determinant 1, The sum of the products of the elements of any row or column with the corresponding cofactors of the matrix is equal to the value of the determinant, 2, The sum of the products of the elements of any row and the cofactors of some other row is zero 3. Let A=(ay) be square matrix of order n, and let Cy be the cofactor of ay then @ [d= aC) (i=1..n) FI (b) lal Yajcy=h n) ja © DhastyCy =Oif ik @) Yay Ca =0 if Fk ‘The determinant of a triangular matrix is the product of its diagonal elements For any two square matrices of the same order, |4B|=IdllB| 6. The value of the determinant of a diagonal matrix is equal to the produet of the elements lying along its principal diagonal 7. The value of the determinant of a unit matrix is always equal to 1 8. The value of the determinant of triangular matrix (Upper or lower) is equal thc product of the elements lying along its principal diagonal 9. The absolute value of a determinant does not change when rows and columns are interchanged, |4|=14" | 10. If any two rows (or two columns) of a determinant are interchanged, the value of the determinant is multiplied by -1 11. If all the clements of one row (one column) of a determinant are multiplied by the some number & , determinant becomes & times of the original determinant 12. If A be an n-rowed square matrix, and & be any scalar, then \xdl=x" Al 13, If two rows (or two columns) of a determinant are identical the value of the determinant is zero 14, If in a determinant each element in any row or column consists of the sum of two terms, then the determinant can be expressed as the sum of two determinant of the same order ips oss i er Hs Kas Near LUT, Ne Dat 016 Ps 1ST, Ca ir "al inialoucadem cou: Werte wndipaadransen ee Scanned with CamScannermae Nee SEL 15, If to the elements of row (or columns) of a determinant added M times 7/1 Own Notes the corresponding elements of another row (or cofumn) the value of the determinant thus obtained is equal to the value of the original —_—_ determinant 16, Let Abe a Square matrix then @ || © Ia © |al=iAl 17: (Adj A) A= A(Adj A)=lAlDy, ll lAl 18. If |4l=0, then adja) 19. adj(AB)=adjB-adjA 20. If the product of two non-zero square matrices is a zero matrix, then both of them must be singular 24. Ifthe matrices A and B commute, then A and B™' also commute ap, det(a-!) =(det 4 | 2.6 Properties of Inverse of a Matrix 1. Every invertible matrix possess a unique inverse 2 (aByt=B at 7. adj A" =(adj ay" If A isa symmetric matrix, then adj 4 is also symmetric If A is symmetric then A™ is also symmetric 10. The adjoint of a diagonal matrix of order 3 is a diagonal matrix a 0 0 mn 0 0 LetA}0 m 0| adj A=|0 an 0 [0 0 7 0 0 am an. det(47!) =(der a) 12. If B is non-singular matrix then the matrix 4 and B~! same determinant Abel ie 13, Ifthe matrices 4 and B commute, then 4”! and B™! also commute ips owe 28 Tin Sara Hans Kos Ne Dai Temallfadlonncademy cow: Wee mw di Scanned with CamScanner15. If a matrix 4 is non-singular, then AB=AC= B=C, where B and are square matrix of the same order as of Properties of Rank of a Matrix 1, Therank ofa matrix 4 is r if t possess the following two properties (@) There is at least one square sub matrix of 4 of order + whose determinant is not equal to zero (b) And all the matrix 4 Counting any square sub matrix of order r+1, then the determinant of every square sub matrix of 4 of order r+1 should be zero. The rank of matrix is the order of any highest order non-vanishing minor of the matrix. The rank of every non-singular matrix of order nism. The rank of a square matrix 4 of order n less than iff is singular ive. [dl=! The rank of every non-zero matrix is 21 The rank of a matrix is r, if there exists r—rowed minor of the matrix which is not equal to zero ‘The rank of a matrix in Echelon form is equal to the number of non-zero rows of the matrix. The rank of the transpose of a matrix is same as that of original matrix, The rank of matrix every element of which is unity is 1 11. Ais a non-zero column and B a non-zero row matrix then rank (4B) 12. Rank ofa matrix is > the rank of every sub matrix 13, Rank of skew-symmetric matrix cannot be one 14. p(A)ene |Al=0 1 1 Let 4=[ay],,.4y €Ror C. Consider 1 oy If n22,p(A)=n-1= p(adja) =f p(A)sn-2= p(adja)=0 18, Let A bean mxn matrix. (a) 0 rank(A) A=0 (©) rank(I,) (@) rank(kA) = rank(A),ifk #0 Scanned with CamScanner2.8 i a { Sea a CaS Ieee ies icant ao tLe ; ————=, 16Let A be an mxn matrix and let Ax=0 for some PE OT Notes x#0¢9rank(A)Sn-1 17, The rank of diagonal matrix equals the number of non-zero diagonal elements it possess but is not true for triangular matrix. 18, Every mxn matrix Aof rank r can be written as A= BCT, where B(mxr) and C(nxr) both have rank r 19. rank(A+B)|rank(A)—rank(B)| 21, rank(AB)< min(rank(A),rank(B)) 22, rank( AB) + rank (BA) (in general) 23. Let 4, bea sub matrix of 4 then rank( A.) ) = rank(B) + rank (C) 27. Consider the matrices Z} then rank (Z;)= rank (A)= rank (CA™"2) if is square and nonsingular rank (22)= rank (D) + rank (BD™ if Dis square and ‘nonsingular, rank Z,)= ran (B)+rank(C), rank (22) rank(B)+ rank(C) 28. Sylvester rank inequality: rank (4B) rank(4) + rank(B) ~ pfor any mx pmatrix Aand pxn matrix B 29. AB =0 implies that rank(A)< p ~rank(B) for any mx p matrix 4 and pxn matrix B Properties of Trace of a Matrix 1. Let and & be square matrices of the same order, and let k.& m are positive scalars, then (®) tr A+B)=0r(A)+ir(B) Scanned witl(b) tr(kA) = kara) ©) trlk.A+mB)=kir(A)+mir(B) @ wla)=(4) (© laa?) =i 44)=Sae 7 2 Forany real matrix 4, tr(44)20, with (UA) <017 A=0 Bole 4 and BB ibe (47a) =r(Ba") tr 4B") =r B74) 4. tr( ABC) = r(CAB) = tr(BCA) but tr(4BC) =tr(ACB) ‘5. For two real matrices 4 and B of the defined as (4,8)= 5,5 ,ayby =r dB mxn matrices, then is not true 1¢ order then Inner product is, (a) (4,8)=(B, A) () (4,B+C)=(4,B)+(4,C) (©) (k4,B) =k(4,B) (@) (4,4)20 with (4,4) =0< 4=0 6. Fora real matrix 4 the Norm of matrix is defined as (44)? lraa~ 55, (@) beall=lal-al (®) |4l20. with 4l-0 if 4-0 © 14+ Blls|Al+|1Bl(oriangular inequlity) 7. Let A and B are square matrix of same of order. AB=T iff B4=1 8 If 4 and D are square, not necessarily of the same order, 4 2B =WrA+trD o(é 0 % Let A and B be mxn matrices, then wr(A"B) = tr BAT) =r 4B") =0r(B7 A) tr( ABC) =1r (CAB) =1r(BCA) bat tr( ABC) =tr ACB) ig not true, 1 2) ,_(-1 0) (2 0 For Example: A=| 4, 4|8=|_) 9 a5 10.1f 4 and D are square, not necessarily of the same order, A OB = trd+ rd o(é 2) 11. Let 4 and B are square matrix of same of order. AB =T iff BA=1 Fin apa Hz Khas, Nese LT, New Dali 016, hs (i) 2653757, Clr BOLING Dies Howse 28 3 Se alualpacdeny com Webule: wwedipucnmresecProperties Important Matrices/ Linear Operators ‘Symmetric Matrices: A square matrix 4 =[ay] is said to be symmetric if A” = A A+A" is symmetric, even if 4 is not symmetric 3. AAT and A™ Aboth will be symmetric matrix. 2. if A and B are symmettic matrices , then (a) AB is not necessarily symmetric (b.) ABis symmetric iff 4.8 commute if A is symmetric, rank(A) = rank(4") A’BA is symmetric if B is symmetric but the converse need not be true jo 3 o1 For Example: Let 4=|) al alt A If A isasymmetric matrix, then adj 4 is also symmetric If A” exist, itis symmetric iff A is symmetric. AP is symmetric when 4 is symmetric. If Aand Bare symmetric matrices of same order then 24: A,B are also symmetric matrices for scalars Ayand 2 Skew Symmetric Matrices : A squate matrix A = aj) is said to be Skew Symmetric if 4 Note: All the diagonal elements are zero. 1. If A bea square real matrix then 4— A? is skew-symmetric 2. Any matrix A can be decomposed as the sum of a symmetric and a skew symmetric matrix uniquely Legal r a(4+4")+ 5(4-47) Drmeirie show aymmetrie . If Aand B are skew-symmetric matrices of same order then, (2) 4B is skew symmettic iff 4B + B4=0 (b) Also 4B ~ BA is always a skew ~symmetric matrix 4A + AyBare also skew- symmetric matrices for scalars 4yand 2p "is symmetric if n is even A"is skew symmetric if nis odd, ‘The value of the determinant of a real skew-s1 symmetric order is always zero, (Note: Provided the field eee is not of characteristic 2), Rank of skew-symmetric matrix cannot be 1 ‘Dips Howe 2, a Sara Haus Kas, Near LT, New DaMLTTOOI, E-mal: nfutaipaendeZom: Webae wrecaec8. Determinant of a skew symmetric matrix =p" ;forsome pe F where F is a field of characteristic not equal to 2. 2.9.3 Idempotent Matrix or Projection Matrix: A square matrix 4 is called idempotent if it satisfies the relation 4? 1. If AB=A and B4=B then 4 and B are idempotent 2. If B is an idempotent matrix, then - B is also idempotent, but B—1 need not be idempotent. Let Aand B be two idempotent matrix of same order, then (A+B) is idempotent iff 48 = BA =0 Let A be an idempotent matrix, then (9) OSTA) sn and THA) eNU0} ©) If TH(A)> mor Tr(4)<0 or THAYENU{O} then A is idempotent (©) Tr(A)=p(A) and (4) =n-Tr(A) never always an idempotent matrix. If T:V(F)>V(F) is a linear operator where dim(V’)=n such that 1? =T then (@ "=PU-T)+p(7) (b) n=n(I-1)+n(7) (© PU-T)=n(T) (@) NU-T)=pT) (©) Image(T ~ 1) = Ker(T) () Image(T) = Ker(T-1) (@) V=Ker(T-1)+Ker(T) (ty V = Ker(T—1)+ Image(T 1) ( V =Image(T) + Ker(T) GV =Image(T) + Image(T- 1) Involutary Matrix or Reflection Matrix: A square matrix 4 is called involutary if it satisfies the relation 4? involutary iff (I+ 4)\—4) =0 ‘] saeRis alvays an involutary matrix. 1 3. If-4 is involutary then (a) A STr(A) Sn, and Tr(A)eZ oe Sart Has Kh New ELT. New Da GFR 1 USTE,CA NOTION ns Ema: lnladioncademcom: Web: wwe dina naee soo Scanned with CamScanner1 Matrices and System of Equations ie If Tr(A)>nor Tr(A)<—n or Tr(A)@Z then A is never involutary. ©) P(A) =n(A-1)-n4 +) © TA)=p(4+1)=p(A-1) (@) n=p(A-D+m4-1) (e—) n=p(A+])+n(A+]) © p(44D=n4-D (@) pA-D=n(4+) (hy) n=nd—DeneD) neTHA) ‘Put Your Own Notes —_—— @ (4-2) If T:V(F)>V(F) , linear operator where dim(V)=n such that T? =I then Image(T 1) = Ker(T +1) Image(T + 1) = Ker(T-1) V =Ker(P-1)+Ker(T +1), V =Ker(T+1)+ Image(T +1) V = Ker(T—1)+ Image(T —1) V = Image(T - 1) + Image(T = 1) Nilpotent Matrix: A square matrix 4 is called a nilpotent matrix if there exists a positive ji i integer k such that A* =O and such least positive integer i such that A¥ =0 is called the index of the nilpotent matrix 4. a ‘| 3@¢ Ris always a nilpotent matrix. Orthogonal Matrix: ‘A square matrix is said to be orthogonal if 447 Example: A= If Ais orthogonal, the following hold: 1, 4 isnon-singular 2 A 3. A? is orthogonal 4. AB is orthogonal when B is orthogonal a i ROR ROSSA % a amarante ose Car Scanned with CamScannerSESE TN eT Ifa matrix 4 is orthogonal, 's rows form an orthonormal set. 7. Ifa matrix A is orthogonal, its columns also form and orthonormal set If rows and columns of a matrix A are orthonormal, then A is orthogonal, 9, IfA and g are non-singular matrices such that 447 = BB™ then there exists an orthogonal matrix P such that 4 = BP 10. If 4 and Bare n-rowed orthogonal matrices. Then AB and BA are also orthogonal matrices Hermitian Matrices: A square matrix A=[ay} is said to be Hermitian if 4° = 4 1, Every diagonal element of Hermitian matrix must be real 2. If 4and B are hermitian matrices of same order, then 42 is also hermitian iff 4B = BA 3. All positive integral powers of are also hermitian 4. If Ais any square matrix, then A+ 4° is a hermitian matrix. 5. Any square matrix can be expressed as 4+iB, where A and B are hermitian matrix. 29.8 Skew Hermitian Matrices: A square matrix 4 =[ay] is said to be skew Hermitian if 4° =—4 The diagonal elements of skew- Hermitian matrix must be purely imaginary numbers or zer0 2, AB- BAis always skew hermitian 3. If A is Hermitian Matrix then iA is skew Hermitian ; 4. If A is a skew-Hermitian matrix, iA is Hermitian 5. If Ais any square matrix , then A—A° is skew hermitian 6. If Ais skew-Hermitian, then diagonal elements of Re(4) are zero, but the diagonal elements of Im(A) are not necessarily zero. 7. Any matrix 4 can be decomposed as the sum of a hermitian and a skew hhermitian matrix Masa)» Manat) Hermition part skew=Hermition part 29.9 Unitary Matrix : =A arvit Ad® A square matrix is said to be unitary if 44 1. If A,B be unitary matrices of same order, then AB and Bdare also unitary own a Sarah Haas Kha, eae LT. New Dh 11016 PL: (i) 3077, Cas ORD Dips Hoe 2 Sl nfoalonacndeon om Webte: nwdiou nora ONS Scanned with CamScanner