CHAPTER 4 Linear Transformations 4.1 Introduction Inthe world of Computer science and Engineering, Mathematics plays a crucial role in solving complex problems, particularly in fields like computer graphics, machine learning, data science, and signal processing. One of the foundational concepts in linear algebra, which underpins many of these applications, is the idea of a linear transformation. A linear transformation is a mathematical function that maps vectors from one vector space to another, preserving the operations of vector addition and scalar multiplication, In simpler terms, it’s a way of transforming data, like moving or scaling points in a coordinate system, while maintaining the inherent structure of the data. For instance, in computer graphics, linear transformations are used to rotate, scale, or translate objects in a scene. In machine learning, they play a crucial role in algorithms like Principal Component Analysis (PCA), where data is projected onto new axes to reduce dimension while preserving as much information as possible. In this section, we will explore the basic concepts of linear transformations, their Properties, providing you with a strong foundation to build upon as you delve deeper into more advanced topics. Let V and WW’ be vector spaces and let T be a function from V to W. Then V is 165166 Engineering Mathematics-2 called the domain of T and WW’ the codomain of T. Let vc V and w « that T(v) = w. Then wis called the image of v under T. The set Of all j v Such vectors in V is called the range of 7. For a given w € W the set of all vin or that T(v) = wis called the preimage of w. Such T:V->wW T(v) range Figure 4.1 Problem 4.1.1. Let v = (1,02) © R’. If: V > W is defined as T(w,,n,) = (er + v2 1. Find the image of v = (1. —3) 2. Find the image of v = (0.0). 3. Find the preimage of w = (1,3). Solution. Given T (v1, v2) = (v1 + v2, 2v2) T(1, -3) = (1+ (—3), 2(-3)) = (-2,-6) T(0.0) = (0,0) To find the preimage of w = (1.3) consider set of alll (vy, v2) such that T(01, 02) = (1,3) (vy + U2, 2v9) = (1,3) rsAen responding coordinates, he a f wutm=1 2u2 = 3 = and vz = 5 So that preimage of w is the set in R? consisting of the 3 — 2 pence 2 13 ingle vector Y= (3 2. Let v = (v1, 02,05) € B®. IFT : V + W is defined as T(o,, 02,05) = | ayy + Uae 202 — ee | (a) Find the image of v = (—4,5,1). ib) Find the preimage of w = (4,1, —1) solution. Given T(01, 02,03) = (2v1 + 02, 202 3uy, m4 ~ v9) (a) Image of v = (—4,5, 1) is given by T(—4,5,1) = (—3,22,—5) (b) To find the preimage of w = (4,1,—1) consider set of all (v.12. v3) such that T (vy, 02.0) = (4.1,-1) (2u, + vg, 2v2 — 3u1, v1 — v3) = (4,1,-1) ®quating the corresponding coordinates, Quy, + =4 Qu, — 3, =1 vs =-1 Consideri rm, we erin 8 the augmented matrix and reducing to row echelon form, we get, 2 1 0/4 1 0 -1;-1 -3 2 0} 1 = ]}01 2]6 1 0 -1}-1 Ong 1) 2 wt168 Engineering Mathematics-2 Solving we get, », = 1, v2 = 2 so that preimage of w js the consisting of the single vector v = (1.2.2). 1 Set inp, XSi ERS eC LOS Let 1" and WW" be vector spaces. The function T : V > W isa linear mation of V’ into Ii” when the following two properties holds for al] and for any scalar c. 1. T(u +v) = T(u) + T(v) 2. T(cu) = ¢ T(u) transfo,. Wvey Pe eee euC kone as A linear transformation T : \' + \ from a vector space V’ into itself is called a linear operator. Problem 4.1.3. Show that the function T(v, v2) = (v) + v2, vy, — 2up) transformation from R? into R?. is a linear Solution. Let u = (1).u2), v = (v1.02) € R® and and let c be any real number Then, using the properties of vector addition and scalar multiplication, we have the following: T (u+v) =T ((ur. ua) +(e, 1%2)) =T(uy +04, uy +t) (Cur + vi) + (ua +02), (ur + v1) — 2(u2 + v2)) = ((uy + ua) + (0) + v2), (uy — Quy) + (vy — 2v2)) = (uy + ug. Uy — Quy) + (vy + vg, V1 — 2v2) =T(uy, u2) +(e, v2) =T(u) +T(v) T(cu) = T(c(u;.u2)) =T(cuy. cup) = (cuy + eur. cu; — 2eu2) = (uy + up. uy — Quy)‘ForGroupA 169 = cT (uy, U2) =cT(u) isa linear transformation. ence : jem 4-1-4 Determine whether the function T : R? — R? defined by T(r. y) = 1) is a linear transformation. tr jution. Let u= (ui, U2), V= (v1, 02) € R sol . T (uty) =T ((u,u2) + (v,0)) =T(u +0, w+) =(u+u.1) T(u) + T(v) = T(uy, v2) + T(r, v2) = (ur, 1) + (ua, 1) = (ur + U2, 2) Here, T (u+v) # T(u) +T(v) Hence T is not a linear transformation on R?. Problem 4.1.5. Determine whether the function T : R? -+ R° defined by T(z. y) = (r’,zy,y*) is a linear transformation. Solution. Let u = (u1,u2), v= (01,02) € R?. T (u+v) =T (uy, v2) + (vr. %)) =T(u + v1, U2 +t) ((ur + 01)?, (ur + 11)(u2 + U2): (ua + v2)°) T(ur, u2) + T(r, v2) = = (ul, uyus, ud) + (v2, Viv2, v3) T(u) + T(v) = O2, uyug + vive, 12 +03) Here, Piuty) ¢ TU) + Te) Hen, ce T i 7 ‘Snot a linear transformation from R? to R°. ~~ eeCheck whether T : Mz. + R defined by T(.A) 1+ Chany te A= : ‘ is a linear transformation. c b, ee a dh eda (lab a2 by oo ( ‘| a [: ‘) _ jartar bi +be “late dtd H=aqtathth+ateatdt+d a by a2 by ravenncr(fea)r(b =a, +b +c, + dy + a2 + bg + c2 + dy Therefore, T(Ay + Av) = T(Ai) | T(A2) rieay=2(« 7 “) a dy _rf{fr o ce. ech = cay + cb) + ec) + cdy a by a dy =e(a, tb) +e. + dh) eT (Ai) -a( = ca, + chy + ec + cdy Therefore, T(cA;) = cT(Aj) Hence T is a linear transformation from Mf... + R. | ag a linea! Problem 4.1.7. Let u € R? and v = (1,1). Check whether T(u) = projet §# transformation from R? to R?. «4= : “For GroupA 171 From Section 3.2.2, proj,u = fn. eR? and let c be a scalar. ) T (cu) = projy(cu) | {cu, v) wv) (u, Vv v) v) (v.v) = € proj, (u) =cT(u) T(u1 + Ua) = projy(u; + uy) — ((u +n), v) wy) _ (uy) , (uv) 2 ( (wv) * = (anv) | (ue.v) ~ (wiv) (vv) = Projyu; + projyu. =T(u,)+T(um) Hence T' is a linear transformation from R? to R?. 41d Properties of Linear Transformations : {s¥a.....¥q} be a basis of a vector space V. Then by Theorem | cc Y€ V’ can be uniquely written as a linear combination of these basis vectors, that, there | exist scalars €1,€2,...,n Such that | V= CLV, + O2V2 tot + CnVne - -hO172 Engineering Mathematics-2 T : V+ V bea linear operator on V. Then by linearity pr, ~ transformations, Peay of, ap T(v) = T(civi + C2¥2 +++ + CnVn) = eT (Mi) + T(V2) +++ + e,T(v,) That is, if {vi,¥a-.,¥n} is a basis of a vector space | T(v,).T(v2)...-.T(¥q) are known, then T(v) can be found for ¥, and ip other words, the set of images of basis vectors of the domain of rend Vey e . determines the transformation. "arsfoma Theorem 4.1.8. Let V and WW be vector spaces and let T be a linear tr, from V’ into 1V, and let u,v € V. Then T satisfies the following prope Terai les, 1. T(0y)) = Ow where Oy and Ow represents the zero vectors of ee V and yy 2. T(-v) = -T(v) 3. T(u-v) =T(u) -T(v) Proof. Letve V. 1. Let Oy: be the zero vector of V. Then, T(Oy) = T(0.v) = O.T(v) = Ow. 2. T(-v) = T(-Lv) = (-1)T(v) = -Tv) 3. T(u—v) = T(u+ (-1)v) = T(u) + T(-v) = Tu) — T(v) Remark 4.1.9. T(y) = Ow is usually written as T(0) = 0 and it is clear that fit is the zero vector in V’, and the second zero is the zero vector in W. Problem 4.1.10. Let T : R — R® be a linear transformation such that 7(1, 01° (3,1,-4), 7(0, 1,0) = (~1,3, -2), (0,0. 1) = (0,2, 1). Find T(1,2,-3)- Solution, Expressing the vector (1,2,—3) as a linear combination of te ———__— ____ForGroupA 173 ‘0,0, 1) we have, 1,0) and ( 0 yo (1,2,-3) = 1(1,0,0) + 2(0,1,0) — 3(0,0,1) therefore, T(1,2, 3) = 1T(1,0,0) + 27(0, 1,0) — 37(0, 0,1) = 1(3,1,-4) + 2(-1,3, -2) ~ 3(0,2,1) =(1,1,-11) 41. Let T: R? — R? be a linear transformation such that T(1,0) = rob ) = (-1,1)- Find T(z, y) and hence evaluate T(1, 4) and T(—2, 1) in em 4. 7(0;1 . solution expressing the vector (x,y) € R? as a linear combination of the vectors d (0, 1), WE ave, 1.0) am (a, y) = (1,0) + y(0, 1) T (x,y) = 2T(1,0) +yT(0,1) =2(1,1) +y(-1,1) =(@—-y,t+y) Hence, T(1,4) = (1— 4,1 + 4) =(-3,5) T(-2,1) = (-2—1,-2+1) = (-3,-1) Problem 4.1.12. Let 7’ : R? — R? be a linear transformation such that T(1,2) = (1,0), T(—1,1) = (0, 1). Find T(x, y) and hence evaluate T(2,0). Solution, Expressing the vector (x,y) € R? as a linear combination of the vectors (1,2) and (1, 1), we have, (w,y) = e1(1,2) + e2(-1, 1D) Corresponding linear system is a-aQ=2 2a +e=y Solving the system, we get, —174 Engineering Mathematics-2 Hence, (1r. 1) Therefore, T(2".y) = oon Hence, (2.00) = (3 -3) Problem 4.1.13. Let T : R* (2.0.-1), T(0,-1.2) = (-3 2° be a linear transformation such that 7(1,1, 1 -1). T(1.0.1) = (1.1.0). Find T(4, 2,0), Solution. Expressing the vector (4.2.0) € Rasa linear combination of the vectors (1.1.1). (0,-1.2) and (1,0, 1) we have, | (4.2.0) = e)(1.1, 1) + e2(0, -1, 2) + ¢3(1,0, 1) Corresponding linear system is ate=4 1 +2q +e Considering the corresponding augmented matrix and reducing it to the row ec lon form, we get,v ForGroupA 175 near system is ow the : M atea=4 -¢2 ~ 64 = —2c3 0, cg = —2, Gs = 480 that pence © (4, 2,0) = 0(1, 1,1) — 2(0, -1,2) + 4(1.0,1) T(4, 2,0) = OT(1, 1.1) — 27(0,-1,2) + 4T(1.0,1) = 0(2,0, =1) ~ 2(-3.2,-1) +4(1, 1.0) = (10,0, 2) plem 4.1.14. Find the linear transformation T : R? — R¥ such that 71.0.0) = a ~1), T(0,1,0) = (2,1, 1), T(0,0. 1) = (1,-1.0). Hence find 7(2, -2.1). < R® as a linear combination of solution. Expressing an arbitrary vector (.r. y. the vectors (1,0, 0), (0,1, 0) and (0.0, 1) we have, (a, yz) = x(1.0.0) + y(0. 1.0) + 2(0,0, 1) Therefore, T(x,y, 2) = xT(1.0.0) + yT(0.1.0) + zT(0.0.1) = x(1,0,-1) + y(2.1,1) + 2(1,-1.0) z+2y+2z,y—2, -zr+y) T(2, —2,1) = (—1,-3, -4) Problem 4.1.15. Find the linear transformation T : R? > Mp, such that T(1,0) = 12 Bae 4 -1 g| and 702, -1) = ie 2] Hence find T(1.1). Solution, Expressing an arbitrary vector (.°,y) € R? as a linear combination of the “ectors (1,0) and (2,—1) we have, (# y) = e\(1,0) + ex(2,-1) a a176 Engineering Mathematics:2 oy + 2g =2 1 =y = =rt2y a= Hence(sr, y) = (a + 2y)(1,0) + (=y)(2,-1) Therefore, T(x, y) = (x + 2y)T(1,0) + (-)T(2,-1) | ee eta -1 of +f af | _ [etsy 2ety =p ay -2y 43 T(1 o-[4 | Remark 4.1.16. A linear transformation is completely described by its action ona basis. = (x +2y) Exercise 4.1. 1. Check whether T : R? + R° defined by T(x,y) = (w@ + y, y, —2) is a liner transformation or not. 2. Check whether T : R® — R® defined by T(x,y, 2) = (c@+1, y-1, z+1)isa linear transformation or not. 3. Check whether 7: R? > My defined by T(x.) = [ ie ry y pa 1}. 4, Check whether T : R! + May defined by T(2.y) = [ity pan ] " linear transformation or not. isd 5. Check whether T ; R? > Mz. defined by T(A) = A? where A € Maas linear transformation. 6. Find the linear transformation T : R’ + R* such that OOuP 177 - r {1 3 fe ET-E T\, 6 d the linear transformation T : R? — M,» such that _ Fin : eee oe |e a 10 70,0= 4 of} 10,1) 9 —7| and T(0.1,0) |i jad the linear transformation T : R® —, R® such that 7(1,1,9) = (1,0,0), 8 Fit 1,8) = Os1,0), 20,3) = (0301). rind the linear transformation T : R° — R® such that T(1,~i,0) = 1 7,1), 70,1.) = (0,1, 1), T(-1, 0,2) = (0,0, 1). Hence find T(1,0,2). 2 ; R? —+ R® is a linear transformation such that T(1,0) = (-1,2.3), T : ra —1) = (1,1, 1), find 7, -1). Answers: 1, Tisa linear transformation. T is not a linear transformation. v T is a linear transformation. . T is not a linear transformation. s T is not a linear transformation. r|*| = a+4b bd] a+ 5b 77a y2) [ez eal 4z—5e 6y+7r : - 1 Pay 2) = (f= + 6y — 22), 5(-3 +3y +22), 5(32—3y+ ) Te.y2) = Qe ty +2, 2e +2, Se+4y +2), T(1,0,2) = (48,7) » T(1,~1)178 Engineering Mathematics-2 4.2 Linear Transformation Given by a Matrix Consider the linear transformation T : R? — R? defined by T (x,y) = (% — 2y,3e + y) We can express this linear transformation using matrices as follows: 1 -2] fr] _ fa-2y TaM=T, 1} tyl > |aety That is. T(v) = Av where v = fl and A= ; al y 301 Ty : fle Recall that the vector (7)..2......,) € R" has the matrix notation |""| | Here we can say that the linear transformation T is generated by the matrix 4, Hence a linear transformations from R" to R” can be generated by Mynx», matrices, This concept may be summarized as the following theorem. Theorem 4.2.1. Let A be an mx n matrix. Then the function T : R" + R™ defined by 7'(v) = Avis a linear transformation from R" to R”. Solution. Let u.v ¢ R". Then T(u+v) = A(u+v) Au+ Av T(u) +T(v) Tv) = A(cv) = (Av) = r=0, y=0 Therefore pre-image of (0, 0) is {(0,0)}- (a) T(1,1) = (-1,-1) : Therefore pre-image of (1,2) is {( Problem 4.2.4. Let T: R’ — R? be a linear transformation defined by 7 ¥) = Ay 110]. . where A = 01 al Find (a) T(1.1,1), (b) the pre-image of (1,3), the Pre-image of (0,0). Solution. Given T(v) = Av where v € B®. Let v = (x,y, 2) pee le Therefore, T(x, y, 2) = oo : That is, T(x,y, 2 (a) TG,2,-1)=(3,3) (b.) Pre-image of (1,3) is {(x,y, 2) € R® such that T(x, y, 2) = (1,3)} That is, (x + y.y — 2) = (1.3) 3 => a=l-ky age of (1,3) is {(1 — k, k, (c.) Pre-image of (0.0) is {(x, y.z) € R* such that T(.r, y, 2) = (0,0)} That is, (x + y.y — z) = (0.0) r+y=0.y- = r=-k y= Therefore pre-image of (0,0) is {(—k. k, k)/k € R}. ; matt Problem 4.2.5. For the linear transformation T : R? > R? given by the i____ For Group A 181 Given T : R? > R? such that T(z, y) = | ( = le = 2 . : solution uy) |br+ay 12a — 5b = 13 7012.8) = 13> ®) => ) p+5a-0 considering the augmented matrix and reducing to row echelon form, 12 -5|13 12 -5| 13 => 5 12,0 0 169] -65 solving the corresponding linear system by back substitution, we get, 12 5 a= 7 b= problem 4.2.6. For the linear transformation T : R? + R’ given by the matrix a=|¢ 2 |, find a and b such that (1,2) = (5,7). 3 a+b Solution, Given T ; R? —> R® such that T(z,y) = [ : Jt 7 3 a+b] ly jae (r+ (a+ b)y|” 71,2) = (5,7) — {ort 5 3+2(a+b)=7 Solving we get,182 Engineering Mathematics-2 4.2.1 Rotation in R° Rotational transformations are crucial in image processing for correcti that have been captured at an angle, thereby making them upright and ima, oriented. This process is particularly important in applications like face = "Oper systems, where the orientation of a face can vary significantly due ¢ in head position. To address this, rotational transformations are a, the face in the image with a pre defined, standard orientation, often, an as a canonical orientation. By ensuring that the face is properly aligneg to feature extraction and recognition processes are conducted, especially in before learning and artificial intelligence applications, these transformations help i ine the accuracy and reliability of the system. This step is essential in minimizin effects of rotational variations, leading to more consistent and accurate recog outcomes across different images. lon RNiti, © diffeng. eee PPlied to \ Example 4.2.7. Find the linear transformation which rotates every vecto, r in &: counterclockwise about the origin through the angle 6. Figure 4.2 : Rotation Solution. Let v = (x.y) € R? and T(v) = T(e,y) = (2’.y/) Let r = |v. z Then cosa =~ => x= ree Por Group A 183 y sina === y=rsing A , os (a +4) = — 1" = rcos(a 49) « r a = 1 (cos a.0084 - sina sin) = Tcosacos8 — rsinasin® = T0088 ~ ysin# sin (a+) =rsin(a +6) y' = r(sin.acos8 + cosasind) =rsina cos + rcosasind = y00s6 + xsing re, serefo uy) [ycos® +2rsin8 Aa | / [ress ysind] cos ~sind| [r © |sin@ cos® ly. refore, the linear transformation T : R? + R? represented by the matrix therefore, : cos — SiN) tates every vector in R? counterclockwise about the origin §* |sind cos through the angle 6. Problem 4.2.8, Find the linear transformation which rotates every vector in 2? 2.8, - counterclockwise about the origin through the angle # = 45°. Solution, Note the result from Example 4.2.7 that the linear transformation i cot snt every vector in TR RE represented bythe matrix A= |g socg | Toates even ® counterclockwise about the origin through the angle #. ee184 Engineering Mathematics-2 Hence the required transformation is represented by the matrix 7 45 -sin4s5]_|Yo ya [z }-\" 2 sin4d5 cos 45 a at v2 V2 The corresponding linear transformation T is given by T(v) = Avwhere y — ; ~My), 7 41 al yal [F]_ |e” ae]? Y=) v2 V2 2 ve. i 1 1 v) T(2,y) = (—r-y. Btrt+s con= (ae ah at va 4.2.2 Projection in R° In Section 3.3.3 of chapter 3, we explored the concept of projection onto subspaces and its applications. In this section, we will examine a basic projection from R® as a linear transformation. Figure 4.3 : Projection onto xy—plane d by Example 4.2.9. Consider the linear transformation 7’: R® -> R* represent= For Group A ~ 185 i a) ri a Ay=|0 10 000 P oR’ is defined by wee T(v) = Av 10 0) fr That is, T(z,y,2) = |0 1 0) |y} =(r,y,9) 00 of [s sider the point (x, y,0) in R° as the point (r, y) in the ry— plane. Here co nally project every vector in R* to zy— plane R?, 1080! sa 4.2.10. Similarly matrices of projections to xz and ye planes are A,. and 4, respectively. 100 000 Az =|0 0 0), A= |0 1 0 001 001 ‘i : ion 4 Figure 4.4: Projection A.. Figure 4.5. : Projection 4, — eee186 Engineering Mathematics-2 Exercise 4.2. 1. Define the linear transformation T : R" > R™ by T(v) = Av w 1-13 2 a —2 4 0 2 |. Find the dimensions of R" and R™, -2 21-6 2. Define the linear transformation T : R" + R™ by T(v) = AV Where As h i" ° 3 Find the dimensions of R" and R”, 3. Let T : R? - R® be a linear transformation defined by T(v) 1 a=|? i . Find (a) T(1, 2), (b) the pre-image of (1,0) = AV Where > (C) the Pre-image of (0.0). 4. Let T : R’ R* bea linear transformation defined by T(v) 1-1 A= |2 0 |. Find (a) 7(3, ol image of (0,00), d) Check whether (1,1, 1) € range(T). = AV where (®) the pre-imaye of (0,2,1), (¢) the pre. 5. Find the linear transformation which rotates every vector in R? counterclock wise about the origin through the angle @ = 30°. Answers: 1. dim(R") = 4, dim(R") = 3 = dim(R") = 4, dim(R™) = 2 o . (a) T(1,2) = (-1.2) (b) Pre-image of (1.0) is (0. -1) (c) Pre-image of (0.0) is (0.0) 4. (a) T(3.2) = (1.6,2) (b) Pre-image of (0).2.1) is (1.1): preimage of (0,0,0) is (0,0) Q pure range(T) wr : ren (FF 43 Kernel and Range ER ome oS inear transformation T : V — W is the set of all vectors » <¢ V. / gernel of 2 7 ped by T into the zero vector in W, denoted by Ker(T). | that are map| Ker(T) = {v€V: T(v) =0} es Vay We Figure 4.6 Remark ie ae By Theorem 4.1.8, if T : V > W isa linear transformation, then ain eat the zero vector in V belongs to the kernel Therefore, the kernel transformation is always non-empty. Problem 4.3.2. Fi hy Find the kernel of the linear transformation 7 : R* + B' defined T (a) .22) = (ay — 2ry, 34y.—5r1) a eeng Mathematics? —_ Solution. To find "er(T), you need t© find all v= (1,72) in R? such that T (ayaa) = (41-280) 32H —52y) = (0,0,0) This leads to the homogeneous system 1-22 =0 > () 3x, = 0 — (it) —51 = 0 — (iii) Solving the system, we get,t1 = and 2 = 0. Therefore, Ker(T) = {(0,0)} = {0} Problem 4.3.3. Find the kernel of the linear transformation T : R* -+ R? define by T(x) = Ax, where s-| 3-3 |) eee Solution. The kernel of T is Ker(T) = {(21,22,23) € R°: T (a1, 22,23) = (0; 0)} 3-3-6 || 2 0 T (x;,%2, 2s) = (0.0) ae anieao [33-3] m|=| 9 To solve the corresponding system of linear equations using Gaussian elimina we start with the augmented matrix that represents the system: a-| 3-3 -8/9 -1 2 3/0 adFor GroupA 189 1 Rak = @ -1 7] -1 2 340 1-1 -2/0 R,+ Ro + RF fe oi aa (A) = 2 is less than the number of unknowns n = 3, so we can choo: jable as arbitrary. Hence the system has non-trivial solutions. Now tht ere rank 21 vari i transformed system 1s 2 — 2 — 23 = 0 — (i) +2 =0— (ii) ay, Substituting a2 = —a in (i), x, = x3. From (ii), 2: z py choosing 73 = t, where ¢ is an arbitrary constant (or parameter). a t 1 a |= |-t} =e] -1 3 t 1 $0, the kernel of T is Ker(T) = {t(1,-1, 1) : t € R} = span{(1,—1, 1)}. (See Figure ??) Rete fy «r(T) is the straight line passing through the origin and the point (1, —1, 1). ee ae aria190 Engineering Mathematics-2 : : Theorem 4.3.4. Let |) and I!’ be two vector spaces. Then the kerne] of the transformation 7°. \" —» It’ is a subspace of V. ‘nea, Proof. From Theorem 4.1.8, 7(0y) = Ow. Therefore, Ov € ker(T) $0 that jy, is a non-empty subset of 1’. (7) Then to show that kcr(7) is a subspace of V” it is enough to show that it is Close under vector addition and scalar multiplication. d Let u.v€ ker(T) and let ¢ be any scalar. Then T(u) = 0 and Tv) = 0. Applying the linearity property of T, T(u+v)=T(u)+T(v)=0+0=0 T(cu) = c(T(u)) = 0 =0 Therefore, u + v € ker(T) and cu € ker(T). Hence ker(T’) is a subspace of V. 9 Pee ee mth kai ad Let T- \’ —+ WW’ bea linear transformation. The dimension of kernel of 7 is called nullity of 7, denoted by v(T). Problem 4.3.5. Find a basis for the kernel and hence evaluate the nullity of the linear transformation defined in the Problem 4.3.3 Solution. In the problem 4.3.3, we have seen that Wer(T) = {t.-1.1): € R} = span{(1,-1,1)} So Ker(T) is @ subspace of 8°, spanned by a vector (1,—1,1). Therefore, ve (U1, =1.1)] is a basis for Ker(T), So the dimension of this kernel is 1 and he nullity of 7 is one Problem 4.3.6. Find the kernel and nullity of the linear transformation ec7 eo R! defined by T(x) = Ax, where x € R' and ForGroupA 191 21 31 0 A- 12 01-1 30-60 3 (Oz0) 0) :2)68 wn, The kernel of T is golutio Ker(T) = { (123,455) € RP: 7 (x1,22,0% = (0,0,0,0)} Bl gn ole 0 12 01-1]|” 0 P(ayteta,t8s) = (06000) = | 3 9 6 9 3) } "> lo oo 02 8ff" 0 rs To solve the system of linear equations using Gaussian elimination, we start with the augmented matrix that represents the system: Rio Ry => Ry > Ry - 2K, Ry Ry + 3K, <—e |12 0 1. 1(5 0® 14 alo 06 -6 3 Ofo 00 0 2 8lo 120 1 1/6 ' 01-1 4 3Io Ree e000 0 @) 4 6 000 2 8/0 12 0 1-1/0 01-13 Z]o ee ee fo 0 0 1 40 00 0 0 0 jo = 3 is less than the number of unknowns n = 5, so we can choose 5 — 3 ~ 2 non leading variables as arbitrary. Hence the system has 2—parameter family of solutions. Now the transformed system is Here rank( A) 1 + 2m +4y—- 25 =0 — (i) 3x2 — 33 +74 — 25 = 0 — (it) ry +425 = 0 —+ (iit) Choosing .r; = s and x; = t as arbitrary, (iit) => xy = —45 = —4t (ii) => 322 = 34g — 24 + 225 = s+2ta ay -2s+t 2 1 ™ s+2t 1 [ 2 x-[m {=| stoefas] ifte] o s,teR 4 Os —4t 0 = 5 Os +t 0 1 | Be {(- 1,1,0,0), (1, 2,0, -4, 1)} is a basis for the kernel of T. 507 the nullity of Ts two. Hen ‘on 4.3.3: Range of a Linear Transformation f a linear transformation T : V — W is the set of vectors in W ‘he range 0! 3 which are images of elements of V. That is, range(T) = {T(v):veV} Figure 4.7 me 4.3.7. The range of a linear transformation T : V > W is a subspace of O06 Brom "Me(P) ig pasha 4.1.8, T(0y) = Ow. Therefore, Ow € range(T) so that Ron-empty subset of 17. Then to show that range(7) is a subspace of be194 Engineering Mathematics-2 WV it is enough to show that itis closed under vector addition ang sla 7 cation. li Let vy.v) € range(T) and let ¢ be any scalar. Then there exist uy. uy € V’ such that T(ui) = "1 and T(u,) = vp, Applying the linearity property of 7, Tu) + Up) = T(mi) + Tus) = vi + V2 Hence v; +¥) € range(T) Pom) = (T(r) = Mi Hence. cv, € range(T) Therefore, range(T) is a subspace of 1". Definition 4.3.4: Rank of a linear transformation Let T. \’ —+ 11’ bea linear transformation. The dimension of range of 7 i called rank of T, denoted by r(7). Basis for the range of a linear transformation: Following two theorems helps us to derive a method to find the basis of the range of a linear transformation. Theorem 4.3.8. Consider the linear transformation T : V -+ W defined by T(v) = Av. Then the range of T is equal to the subspace spanned by the column vectors of A (ie, the column space of A). Theorem 4.3.9. Let A be an m x n matrix. Then the columns of A corresponding to the columns of leading variables of its row echelon form constitute a basis for the column space of A. Steps to find the basis for the range of a linear transformation: 1. Write the given linear transformation in the form T(v) = AV. 2. Reduce the matrix 4 to its row echelon form R. a_ForGroupA 195 ote the columns of R corresponding to the leading variables, 3, Ne 4, Form the basis for the range of T’ using the corresponding columns of A. m 4.3.10. Find the range and rank of the linear transformation ple mots RY defined by T(x) = Ax, where x ¢ R° and 31 0 1 01-1 30-60 3 00 02 8 | 21 31 0 12 01-1 solution. Given T(x) = Ax where A = . oa 00 02 8 Asin Problem 4.3.6, reduce A to its row echelon form to get, @ 20 1 1 0@-1 4-4 0 00@M4 0:0 0 6 0 Here the leading variables in the row echelon form corresponds to the first, second and forth columns of A. Therefore a basis for the range of T is {(2,1, -3,0), (1,2,0,0), (1, 1,0,2)} Hence Rank of T= 3. Problem 4.3.11. Find a basis for the range and rank of the linear transformation R' — R® defined by T(x) = Ax, where x € R' and 11 0 3: A= 1 1-3 -1 -2 0 3-2 Nee156 Engineering Mathematics2 Ax where A = Solution. Given T(x) Reduce .{ to its row echelon form to get, @M-1 0 3 R+R-R _, |g 9 3 _4 Ry > Rs + 2Ri ooo 3 4 tl 0s Ry R+R, = |0 @ -3 -4 0 0 0 Q Hence the row echelon form of A is @® o 38 R= 0 2 -4 0 0 0 Here the leading variables in the row echelon form corresponds to the first and second columns of A. Therefore a basis for the range of T is {(1,1,—2),(-1,1,0)} Hence Rank of T= 2. Theorem 4.3.12. (Sum of Rank and Nullity) Let 7: V > W be a linear trans: formation from an n-dimensional vector space V into a vector space W. Then the sum of the dimensions of the range of T and kemel of T' is equal to the dimension of the domain V’. That is, Dimension of range(T) + Dimension of ker(T) = Dimension of V. That is, Rank(T) + Nullity(T) = n Note: Above theorem is known as Rank-Nullity Theorem. Problem 4.3.13. Let T : R° —+ R® be a linear transformation. eeee dimension of the kernel of T when the dimension of the range is 3- @ rind the ) rind the rank of T when the nullity of 7 is 2. 0 (o) Find the rank of T when ker(T) = {0}. d we will use the result Rank(T) + Nullity(7) = dimension(R®)=5 — (1) dimension of range(T) = 3. That is, Rank(T)=3. (2) = 3+Nullity(7)=5 ‘Therefore, Dimension of the kernel of T= Nullity(T)=2. Given, nullity of T is 2. gotution. Here ive, ® (1) => Rank(T)+2=5 Therefore, Rank of T=3. io iven, ker(T) = {0}. Therefore, by Remark 2.4.11, Nullty(T)=0. (2) = Rank(T)+0=5 Therefore, Rank of T=5. problem 4.3.14. The linear transformation T : R? + R® by T(x) = Ax where 5 -3 A=]1 1 -1 Find (a) ker(T), (b) nullity(7), (c) range(T), and (d) rank(T). Also verify that tank(T)+nullity(T) =dim (RR?) . Solution. Given T : R? -+ R* defined by T(x) = Ax. The kernel of T is {(x, y) € R? : T(x, y) = (0,0,0)}. 5 3 T(w,y)=|1 1 fl - Cor i “sponding linear system is " oes198 Engineering Mathematics-2_—_ xty=0 ™~ r-y=0 5-30 Considering the augmented matrix A = 110 and reducing itto roy 1-10 on form, we get, 5-30 110 =|1 1 0) => J010 1-190 000 Corresponding reduced system is r+y=0 y=0 Solving the system, we get « = 0, y= 0. Hence ker(T) = {(0,0)} and nullity(T)=0. Now, consider the matrix related to the linear transformation T and its row echelon rm, 5-3 i A=]1 1] => ]01 1-1 00 Here the first and second columns corresponds to the leading variables and hence the first and second columns of A constitute a basis for range(T) given by {(5.1.1), (-3,1,-1)} so that Rank(T)=2. Also rank(T) +nullity(T) = 2+0 = 2 = dim(R?). Se co ee Sercses For the following linear transformations 7: R” —» R® defined by 7(@) = 4" find the following: (@) Basis for the range of 7. (b) Rank of 7. CC| (CBasis for the nullspace of T, @ Nullity of T. 2 4 ae LTR OR, A I ‘| a aT: RR, A=} | | roa a T:ROR, A=|-5 3 : i 1 | 101 O10 101 4.7: ROR, --| 5. 1:R°>R, --| Answers: 1, (a) Basis for range(T) = {(2,—3),4,—4)}, (b) rank(T)=2, (Nullspace={(0,0)}, (4) Nullity=0 (a) Basis for range(T’) = {(1,0),(—1,1)}, (b) rank(7)=2, (Basis for nullspace= {(—4,—2,1)}, (d) Nullity=1 . (a) Basis for range(T) = {(4,0,1),(0,4,-1)}, (db) rank(T)=2, (© Nullspace={(0,0)}, (d) Nullity=0 (@) Basis for range(T) = {(1,0,1), (0,1,0)}, (b) rank(T)=2, (OBasis for nullspace={(—1,0,1)}, (d) Nullity=1 a) Basis for range(T) = {(2,—2,1)}, (b) rank(T)=1, (Basis for nullspace={(1,1,0),(1.0,—2)}, (4) Nullity=2 ee200 Engineering Mathematics-2 4.4 Matrix of a Linear Transformation Matrix representation of a linear transformation offers several advant; a Gilly in the context of computational linear algebra and related fede °°? transformations such as rotations, scaling, and translations are easily eed inear ing matrix operations, making matrix representation essential in graphics. an Us. like data science and machine learning, matrices are used to represent and m, fields late datasets, where linear transformations help in tasks like dimensionality rat tion Overall, matrix representation of linear transformations provides a poweer flexible. and widely applicable framework that simplifies both theoretical ae and practical computation. Definition 4.4.1: Standard Matrix for a Linear Biscuteuciscos Let TR" + R” bea linear transformation and B = {€1,€2,...,€n} be the standard basis for R". Then standard matrix for a linear transformation is defined by where T(e) = Amy m2 Grn Ami for 1 = 1,2. n. Note that T(v) = Av, ¥ VER". Problem 4.4.1. Find the standard matrix for the linear transformation T . R° — R? defined by (x — 2y.32 + y— is for Solution. Let 4e, = | 0 | .e2=]| 1 |,es=| 0 | ? be the standard basseee ror croup Ae 201 fg 7(e1:) =T0,0,0) = (1,3) => Tle)yo7 ‘ 7 0 7 (ex) = T(0,1,0) = (—2.1) => Tle) = 1 0 7 (ea) = 20.01) = (1) => Tes) =7 (|: I: columns of standard matrix A consist of T (e,) ,T (ez), and T (es), and we have A=[T(e) Tee) rs ]=[} . ‘| ‘the r | teark: To verify that Av = To, let v= [x y 2)" and consider, | ho) Olle r—2y | a=) 1 If [25% ]-20 Problem 4.4.2, Let 7’: R? +> R? be a linear transformation defined by T (x,y) = (w+ y,22 -y) Find the standard matrix for T. Soluti 1 ae {* : o | e= [: } be the standard basis for 22 Pe) =TA,0)= (1,2) ra)=7((}}) -|:|202 _ t Te) = T(0.1) = (-2.1) ~reo=7([t])-| ' . ~1 The columns of standard matrix A consist of T (e,).T (e2), and we have T(e) Tle) |= [: i] Problem 4.4.3. Use the standard matrix for the linear transformation 7 R: defined by ?R sp T (x,y, 2) = (2x + y.3y — 2) to find the image of the vector v = (0, 1.—1). 1 0 0 Solution. Let ¢e; = | 0 }.e2=| 1] ,es= | 0 be the standard basis for 0 0 1 R*. T (e,) =T(1,0,0) = (2,0) => T(e) = oon tl T (e2) = T(0.1,0) = (1.3) => T(e2) = ; -[3] 13 0 0 0 T (es) = 7(0.0.1) =(0.-1) => T(es)=T | | 0 -| 1 i The columns of standard matrix A consist of T(e1) .T (ea). and T (es), and were A=|T(e) T(ex) rien |=[§ 3 —eee Rant r al 1 -1 and we know that, i rmoarlis sity grerefore T(0,1,—1) = (1,4). problem. 4.4.4. Find T(1,—5,2) by using the standard matrix for the linear trans. formation T = R* + R*, defined by T(z, y, 2) = (2¢,x + yy +z,0 42) i 0 0 Solution. Let )€:= | 0 |, @2.=| 1), es=| 0 | } be the standard basis for 0 0 1 BR. 1 T (ex) = T(1,0,0) = (2,1,0,1) 0 => Tla)= ° —— q 0 T(e) = 7(0,1,0) = (0,1,1,0) => T(e)= ‘{ | 0 , 0 (©s) = 7(0,0,1) = (0,0,1,1) — T(e)=T\)0\\= me oS Ye i _sts—ststw304 Engineering Mathematics-2 — The columns of standard matrix A consist of T (e1) ,T (e), and 7 >. 3), ang We bag 4=[Tie) Te) Te) | = HON orreo HrHoOO T letv=|1 -5 2 and we know that, Therefore T(1.—5.2) Definition 4.4.2: Matrix of a Linear Transformation Let T be a linear transformation from an .— dimensional vector space V into an m—dimensional vector space W’, and let B = {vi,v2,..., Va} be a basis for and C = {w,.wy.....Wm} be a basis for W. Then the matrix of 7 with respect to 8 and C is defined by an ay, 422 +++ On iyo- |e where [T(v)]_.= Amy Om Ginn Oe. which is the coordinate matrix of T(v,) relative to the basis C for ' 1,2 n 0) Problem 4.4.5. Let 7: R? + B? be a linear transformation defined by 7(" (x + y. 2x — y). Find the matrix for T relative to the bases B = {(1,2).(=1,1)} and C = {(1,0), 0, D} —_jon. Let v) = (1,2), ¥2 = (~1, 1) and wy = (1,0), wa = (0,1). finition of T, we have gol py the del T (vi) = T(1,2) = (3,0) = 3w; + Ow2 T (v2) = T(—1,1) = (0,3) = Ow; — 3w2 The coordinate matrices for T (v,) and T (v,) relative to the basis C’ are TwWle= [3] and {I (va)],.= [ 5 tence the matrix for T relative to the bases B and C is given by, c_|3 0 me-(3 9 Remark 4.4.6. Comparing with Problem 4.4.2, we can see that the matrix linear transformation may be different with respect to the bases that we choose Problem 4.4.7. Let T : R? > R? be a linear transformation defined by T(r. =) — (3r — 22, 2y — z). Find the matrix for T relative to the bases B = {(1,0,1), (1,-1,0), (0, 1,1)} and C = {(1,1),(1.0)} oe Let v. = (1,0,1),v2 = (1,-1,0),vs = (0,1,1) and w, = (1.1),w: = 0). By the definition of T, we have T (m1) = T(1, 0,1) = (1,-1) = -1(1,1) +2(1.0) = —tw: + 2we P (va) = T(1,-1,0) = (8,2) = -2(1,1) +5(1,0 Tv) 2w, + SW2 = T(,1,1) = (~2,1) = 1(1,1) - 3(1,0) = Iw, — 3we ei eee206 Engineering Mathematics-2 _ The coordinate matrices for T (vi) 7 (v2) and Ts) relative to the basis ¢ 3, ¢ A IT (walle = [7]. {T (va) and C is the matrix formed by taki NS thes se IT (1) Now the matrix for T relative to B coordinate matrices as columns. That is, ofa 0 Tr=| 9 5 -3 Theorem 4.4.8. Let 7 be a linear transformation from an n— dimensional yee, r mal vector space W, and let B= {Vi,¥V2....,¥4) be Wr} be a basis for W. Then a [ro].- Ela Me Problem 4.4.9. For the linear transformation T R? > R? in Problem 4.4.5, use the matrix of T to find T(v), where v= (5,4). Solution. From Problem 4.4.2, B = {(1,2),(-1, D}, C= {(1,0), (0, 1)} and [T]p = 3.0 0-3 Now, v= (5.4) = 3(1.2) —2(-1,1), which implies space |" into an m—dimensio basis for \" and C = {wi.Ws So by Theorem 4.4.8 TWw)e = [T\Slv\n = : 5 [2] That is. [T(v)\¢ a 6_207 For Group A_ gince C= (1.0), (0,0), T(v) = 9(1,0) + 6(0, 1) = (9,6) problem 4.4.10. Let B= {(—3,2),(4,—2)} and © = {(-1,2), (2, -2)} be the bases 2 . and let |_|, ;|be the matrix for T : R? 5 R? relative to B and C. Find q(v), where v= (11, -6). for’, solution. Given B = {(—3, 2), (4, —2)}. y= (11,6) = —1(-3, 2) + 2(4, ~2), which implies so by Theorem 4.4.8 ITMle = (Tis w= [2 7) f-2]_ [6 ee | -a7 2\~ ha Given that, C = {(—1, 2), (2, —2)}, so it follows that T(v) = 16(~1, 2) + 17(2, =2) = (—16,32) + (84, 34) = (18, -2) Problem 4.4.11, The matrix representation for T : R? — R’ is Tl = {| where B = {(1, 1), (2.1)}. Find T(x, y). Solution, To fir We need to a 'nthe basis ind the image of (:r; y) under the linear transformation T : R? ply the matrix representation of T to the coordinate vector of (.r. ) B= ((1,1),(2,1)}. Step 1. i 5 ) P 1: We want to express (1”, y) as a linear combination of the basis vectors (1. 1) tm208 Engineering Mathematics-2 and (2,1). That is, we seek coefficients a and b such that: (x,y) = a(1, 1) + (2,1) This gives us the system of equations: w=a+2 y=atb From the second equation y = a +6, we geta = y —b. Substitute this into the first equation a=(y—b)+2b=y+b => b=a-y a=y—b = a=y-(e-y)= 2-2 : Qy-x So, the coordinate matrix of (sr, y) in the basis B is, [(x,y)]a = a _ ‘] Step 2: So by Theorem 4.4.8 Te, wla = TB vle = [5 ‘| [7-3] _ | 2y-2+22-2y | ~ |6y — 3a + 4x — 4y _f[« © [Qyta Step 3: Given that, B = {(1, 1), (2,1)}, so it follows that T(x, y) = x(1,1) + (2y + x)(2,1) = (c+ 2(2y +2), r+ (2y +2) = (3x + dy, 2a + 2y) Problem 4.4.12. Let B = {(2,0,1), (0,2, 1), (1,2, 1)} and_ ForGroupA 209 1 = {(1,0,0,1), (0,1,0, 1), (1.0, 1,0), (1, 1,0, 0)} be the bases for R® and R‘ respec- tively. Find T6, 0, 2) for the linear transformation defined in Problem 4.4.4, by ysing the matrix of linear transformation related to B and B’. solution. Given B= {(2,0,1), (0,2, 1), (1,2, 1)} B’ = {(1,0,0,1), (0,1,0, 1), (1,0, 1.0), (1, 1,0,0)} so find the matrix of T(x, y,z) = (2a,0+y,ytz,0+2) relative to the bases B and B’: step 1: Apply T to each vector in B = { v1,¥2,¥s}- T(2,0,1) = (4,2,1,3), T(0,2,1) = (0,2,3, 1), T(1,2,1) = (2,3,3,2). Step 2: Express T(v:) in terms of the basis B’. (4,2,1,3) = 2(1, 0,0, 1) + 1(0,1,0, 1) + 1(2,0,1,0) + 1(1,1,0,0), (0,2,3, 1) = —2(1, 0,0, 1) + 3(0,1,0, 1) + 3(1,0,1,0) — 1(1,1,0,0), (2,3,3,2) = —1(1, 0,0, 1) + 3(0, 1,0, 1) + 3(1,0,1,0) + 0(1, 1,0,0). The coordinate matrices for T (v1), T (v2) and T (vs) relative to the basis C are: 2 2 -1 iwle=| 1]. eedle=| 5). Pole=| 5 1 -1 0 Step 3: Form the matrix for T relative to B and B’ by using these coordinate210 _Engineering Mathematics-2 matrices as columns 2-2 -1 w_|1l 3 3 T=), 3 3 1-1 0 We need to express the vector v = (5,0, 2) as a linear combination of the vectors in B: (5,0, 2) = a(2,0, 1) + 6(0, 2, 1) + (1,2, 1). This gives us the system of linear equations: 2a+e=5, 2b + 2c = 0, atb+c=2. From the second equation: 2b+2c=0 > Substitute b = —c into the third equation: a-ct+c=2 > a=2. Substitute a = 2 into the first equation: 22) +e= = 44+c¢=5 = Using c= 1 in b= b -1. Thus, the coordinates of (5, 0, 2) relative to the basis B are: fvla= | -1ForGroupA 211 go by Theorem 4.4.8 2-2 -1 ' 5 Tl imp 1 3 3 2 (T)]|e = (T18 Wv\p = _ (T] ve ar ; 1-1 0 3 Given that, B’ = {(1,0,0,1), (0,1,0,1), (1,0, 1,0), (1, 1,0,0)}, so it follows that T(v) = 5(1, 0,0, 1) + 2(0, 1,0, 1) + 2(1, 0, 1,0) + 3(1, 1,0,0) = (10,5, 2,7) Remark: Above problems demonstrate that, to find the image of a vector under the linear transformation 7’, we need only the matrix of the transformation. This approach is especially useful in computational applications because matrix multi- plication is well-suited for implementation in machines. Exercise 4.4. 1. Find the standard matrix for the linear transformation T : RS + R? defined by T(x, y, =) = (a — 2y, 2x + y). 2. Find the standard matrix for the linear transformation ab a+2b+3c wb -3ce+4d T: i = Max2 > Mea defined by T : Sla|o ap seo 3. Use the standard matrix for the linear transformation T : R’ + R® defined by T(a,y) = (a — By, 20 + yy) to find the image of the vector v = (—2, 4). 4. Use the standard matrix for the linear transformation T : R* — R* defined by Qe + By — 2,32 — —yts) T(x.y.2) to find the image of the vector v = (1,2, —1)- - For the linear transformation 7’: R? — R? in Problem 4.4.5, use the matrixOf T to find T(v), where v = (2,1). 6. Let T ; R? + R® be a linear transformation defined by Tay). (2a, x + y, 3y). Find the matrix for T relative to the bases B={(1,1),(1,-1)} and C = {(0,0, 1), (0,1, 0), (1,0,0)} 7. Let B= {(2,0,1), (0,2,1), (1,2, )} and C = {(1,1,1), (1,1,0), (0,1, 1)} be the bases for R*. Use the matrix of linear transformation related to B and ¢ to find T(4,—5, 10), for the linear transformation T : R° — R°® defined by T(z,y,2z) = (@+y + z,22 — 2, 2y —2). Answers 1-2 0 1. A=[T(e) Tle) reo }=[} : °| 1 2 3 i) 2.A= o 2 -3 4 3-4 0 -5 00 0 0 3. T(—2,4) = (-14,0,4) 4. T(1,2,-1) = (9,5,-1) 5. T(2,1) = 3(1,0) + 3(0, 1) = (3,3) eI 3-3 =/2 0 2 2 7. T(4,—5,10) = (9, 16, —20)