4 Matrices 4.1 DEFINITION Let us consider a set of simultaneous equations, X+2yt3r457=0 4xt2y4+52471=0 1 the coefficients of x, y, z fof the above equations and enclose, 3xt+4y+22461=0. Now we write dow within brackets and then we get 1235 mei e257 426 + woe ‘ows and columns , The above system of numbers, arranged ini a rectangular array in 1 MNS ang bounded by the brackets, is called a matrix. ents. It is termed as 3, It has got 3 rows and 4 columns and in all 3x4 a 12 ae the fit ae x matrix, to be read as [3 by 4 matrix]. In the double subscripts of an elen cou the FES subi determines the row and the second subscript determines the column in which the element lies, 4, lies in the ith row and jth column. 4.2. VARIOUS TYPES OF MATRICES (1) Row Matrix. If a matrix has only one row and any number of columns, it is called a Row matrix, e.g., (2739) (2) Column Matrix. A matrix, having one column and any number of rows, is called 1 Column matrix, e.g.,| 2 3 + in which all the elements are zeros, (3) Null Matrix or Zero Matrix. Any matrix called a Zero matrix or Null matrix e.¢., 0000 0000 ich the number of rows is equal to the numbe (d) Square Matrix, A matrix, in whi columns, is called a Square matrix e.g., 25 14 A square matrix is called a diagonal matrix (e) Diagonal Matrix, : : A elements are zero e.g., » if all its non-dia 196matrices 197 loo o 30 (p Unit or Identity Matrixze eo 4 jemeats aire unity and non- ahg hob f (i) Skew Symmetric Matri Atrix. A squaré matrix is called skew symmetric matrix, if (1) ay = ~ 4% for all values of i andj, of A’ (2) All diagonal elements are zero, e.g, ie + eBay 0 -h ~g ho oO -F get 0) i) Triangular Matri: . fe is called an atrix. A square matrix, all of whose elements below the leading diagonal ae easiness eee ae elements above the leading diagonal are zero, is called a lower triangular matrix e.g., feo 041 a0 006 5 6:77 Upper triangular matrix Lower triangular matrix 4 the . If in a given matrix A, we interchange the rows an‘ he transpose of the matrix A and (j) Transpose of a Matri corresponding columns, the new matrix obtained is called is denoted by A’ or AT e.g., 234 216 A=|1 0 5|.4,-\3 0 2 678 458 (&) Orthogonal Matrix. A square matrix A is called an orthogonal matrix if the product of the matrix A and the transpose matrix A’ is an identity matrix e.g., A.A =T if |A| = 1, matrix A is proper. (0) Conjugate of a Matrix Lf iti 2-38 4 ae a=[ Nn -i 51 Conjugate of matrix of A is A = _f iri 2438 4 a=(15% i seve (m) Matrix A®. Transpose of the conjugate of a matrix A is denoted by A°.198 a, aoe 4 Let efit, 223! ate pot 2438 4 7-21 +h 342F 1-i0 7-24 (Ay =| 2434 i 4 342 1-i 7-2 ~ Aeaf2+3r 1 4 3428 a (n) Unitary Matrix. A square matrix A is said f0 be unitary if APA =T Iti -14i e8. 2 0 , AAP aT l+i _ 4 2 2 __,, (©) Hermitian Matrix. A square matrix A = (ay) is called Hermiatian matrix, if ey... #yth element of A is equal to conjugate complex j-ith element of A. In other words ay = By 1 243i 3+i ek 2-31 2 1-27 3-1 1+2i 5 Necessary and sufficient condition for a matrix A to be Hermitian is that A = 4® ie, conjugate transpose of A or A=). () Skew Hermitian Matrix. A square matrix A = (aj) will be called a Skew Hermitian matrix if every i-jth element of A is equal to negative conjugate complex of j-ith element of A. In other words ay = -% All the elements in the principal diagonal will be of the form ay = Gy or ay +% = 0 If ay =a+ib then %; = a-ib (a+ ib)+(a-ib)=0 or 2a = Oora =0 So ay; is pure imaginary or ay = 0. Hence all the diagonal elements of a Skew Hermitian Matrix are either zeros or pure imaginary. i 2-37 4453 eg. -(2+3) 0 2i -@-Si) 2i -3F The necessary and sufficient condition for a matrix A to be Skew Hermitian is thataut matrices : tan * 3 tan 80d 18 a unit 2 ‘| MU matrix, show that 4A = anal Sne = Jones is pete imc. wae[ core sina & ~sin ct, cos a | *EN show that 4s j ot 7 = ven tow haar = [+28 — n 12m | 10. WF) = 9 -20x+8, fing ] , fim 1 | YAywhere aa} 1 3 3 | -2 -4 4 show that[ 050 ~sino) | 1 ~tan& ' | ane co|7 e “nz || t one tan 1]]-wn® 1 ] Ss -=-3 4 [ | Ae|2 12 7 =; 4 'show that A? = 471, } 43. Verify whether the followi i i. lowing matrix is orthogonal. A = 4| -12 3 22 } Verify th 24.2 { 14. Verify that 3/-2 1 . 2 [is an orthogonal matrix. -2 -2 -1 __ 60s} 0 sing 15, Show that] sin@sin@ cos@ ——sin @cos | is an orthogonal matrix. —cos@sin@ sin@ cos @cos cos 8 Show that A = or’ 0 -sin@ 0 cos® 17. IfA and B are square matrices of the same order, explain in general (@ (At BP # A?+2AB+B? (i) (A-By? # A?-2AB+ BP Let A and B be any two matrices such that AB = 0 and A is non-singular. (b) Bis also non-singular; (c) B= A; (@ Bis singular. (AMLE.TE., Winter 1996) Ans. 0 sin is an orthogonal matrix. 16. (ii) (A+B) (A-B) 4 AB 18. Then (a) B= 0; 4.10 ADJOINT OF A SQUARE MATRIX Let the determinant of the square matrix A be | Al. a a 43 by ba bs a 4 43 Then|A| = co If Az|b b bs ici 2 The matrix formed by the co-factors of the elements in A, Az As B, Br Bs aa G [Al isa" 212 Malticog : = = byes bye where Ay © [- we byey ~ bea» A: | ay Os] 2 - ayes + ane oe ay @] 2 -ayer + 4201 ne I: 2| eaernae, BET la “| 1c 6 ay a3 | 2 -ayby + asbr a= a ml = bya, C=] bs 2! 22 |e ab: o= [2 | -ennan ‘Then the transpose of the matrix of co-factors A By OL Az B, G Ay Bs Gs is called the adjoint of the matrix A and is written as adj A- 4.11. PROPERTY OF ADJOINT MATRIX ite on djoint is equal to unit matrix multiplied 2! e determinant The product of a matrix A and its a = A-(AdjointA) = JAL-2 A If A be a square matrix, then (Adjoint A) -A a a 4% Ay Bi Ot Let A=| 1 2 &3 | and adj.A = Az B, © ca As Bs. Cs ay, a2) OT pat BY a A-(adj.A) =| 21 22 mh Ba a] a 3) [As BG aA, + @ Ant as As a, By + a2 Bz + 3 Bs ay Cy + a2 Cp +4303 =| by Ay + br Art 343 by By + bz Ba + b3 Bs by Cy + b2 C2 + 3s cy Ay +0242 teas cy By + C2 Ba + 63 Bs cy Cy +02 2.3 C3 JA] 0 0 100 oat 0 f=lalpo'l oO = |All 0 0 [Al 001 1.12 INVERSE OF A MATRIX If A and B are two square matrices of the same AB = BA=1 and A is the inverse of B. order, such that (7 = unit matrix) 1en B is called the inverse of A ie. B = AW! Condition for a square matrix A to possess an inverse is that mai 2, A] #0 If A is a square matrix and B be its inverse, Taking determinant of both sides [AB] =| 1] or [ANB] = 7 trix A is non-singular, then AB=7/> aatrices 213 From this relation it is clear the matrix A is non-singuta, nt !41 #0 r. ie. spo find the Inverse matrix by the he we know that A - (Adj 4) of adotm mane HA) = Ay A: (Adj). [Ay Provided |A| # 0 or and hin Act 3 a) jap ae”) Example 18. If A =| 2 -3 4 G. a7 4 ind ay . 3-3 4 Solution. A=|2 -3 4 oO -11 [4] = 3(- o CS eee 3-4) (3-0) (3-0) (12412) (1248) (-9+6) ‘Therefore the matrix formed by the co-factors of [A | is 1-2 -2 1-1 -1 3 3 |,Aaqg.A =] -2 5 0 -4 -3 ics ee if 1-0 1-1 0 Ala TypadaA=y[-2 3 -4]/=] 72 3 -4 Ans. -2 3-3] [-2 3 -3 4 7 |, prove that A~! = A’, A’ being the transpose of A. (AM.LE., Winter 2000) 4 -8 4 1 Solution. If ' 14 AA 9 4 47 1-8 4 af Stilo 81 -8-8+16 81 oO oO = o 81 | 0 o 8! r Proved— 214 that or or and or Matti, pve that A! A satisfies a relation A?+A— ie Example 20. If a matrix Av! = +A, 1 being an identity matrix. » fc Atn=1 Solution, Here A?+A-/ ori AreAria EOF A(At) lataer= dl | g [A] # Oand so AW! exists. Again AM+A-1=0 or AP+A . ~~) Multiplying (1) by A~! we get AvM(A24A) = ATH or AFT = a Al SI+A Proved. , rder then, Example 21. If A and B are non-singular matrices of the same 07 ee " (aby! = tA (AMLE., Winter 1998, Summer 1996 Hence prove that (A-!)" = (A")-!_ for any positive integer ™- Proof. We know that A (ape-'}-A} = [A (BB) A = [AN -A71 = AAT = 7 <1A-1 (4B) = B-NA-t (AB = BOA AYP Also B = Bo By= BB =I -14-! is inverse of AB. (AB) - (B-'A- By definition of the inverse of a matrix then B Proved. BAT! = (ABy! . (Amp = facamipt = amy at gee alle (eee a geen, cam=3yh ArH (AH? = (Am 3) “1 a = A-amayt (AH satay = ayn Prove that the inverse of a matrix is unique. of a given matrix say A. Proved. Example 22. Proof. We suppose that B and C are two inverse matrices Then AB = BA=1 “Bis inverse of A. AC =CA=1 + Cis inverse of A. But C-(AB) = (CA)-B (Associative law) -B or C=B : Proved. cl Hence the inverse of matrix A is unique. Example 23. Find A satisfying the Matrix equation. Bab? 3)-[3 <1] swion [2 3]4f-2 2] -[-3 4] Both sides of the equation are pre-multiplied by the inverse of [2 ! ] ie. [ a2 | [3K SRL as wr 3 2masse? [o thr: le 2 a[-3 3]- -7 again both sides are post-mutti (3318 3]-[2 lo A = 24 [3 ra [-34 A= 24 M4 Example 24. Given As 2 2 find C such thar ca =[h Solution. BcA=| 4 [o Ta 23 7 Cl 4 ie [3 4] E 3 | Le 1a ce} 2. 4.1 |= [2 \ 231 = raid «fz tle 231 1 lr or Gia --[- 23 1 c=-[ ro HO HO WAM RO 9 -14 9 ~14 9 -14 ] | plied by the inverse of le I | 215 Al1 2 322 8 Grenson a= [3 3 i] 4 342 Compute det (A), A~! and the matrix B such that AB =| 1 6 il sa 56 il Also compute BA. Is AB = BA? 9-2 -4 100 an. 5,2] 1 2 -1,a=|0 2 04Be aa ia 7, a 9. Find the condition of k such that the matrix 134 A=! 3 k 6/has an inverse. Obtain A~ for is f (AM.LE.TE., Summer 1997) 3 1 -29 17:14 Ans, k#-%,A'=5) -9 5 6 > 8) 16 -8 -8 10. Prove that (A~!)7 = (A7)-!. 11. Let [be the unit matrix of order n and adj. (21) = 2 1. Then & equals @i (6) 2 (c) n-1 (d) n. Ans. (c) 4.13, SOLUTION OF SIMULTANEOUS LINEAR EQUATIONS Let the equations be axtanytajsz=d b xt byy+byz = dy cyxtaytesz We write the above equations in the matrix form a xtay+a3z d, a a a3] x dy dxthytbsz} =| d| or |b) b by |) y|=| d CX+CQV+C3z 43 Cy C2 cy fz dy AX=B (1)250 4.25 CHARACTERISTIC ROOTS OR EIGEN VALUES (a) For a given square matrix A, A —A/ matrix is calle A. is scalar and J is the unit matrix, 22 1 Let A=/1 3 1 12 2 22 100 2-2 2 1 131]-aAj/o10/=/1 3-- 1 122 ool 1 2 2-2 characteristic matrix d the characteristic matrix, where (6) Characteristic Polynomial The determinant |A —A/| when expanded wil characteristic polynomial of matrix A. 11 give a polynomial, which we call as 2-4 2 1 For example | 1 3-A 1 1 2 2-2 = (2-A) (6-SA+A2-2)-2(2-A-1) +1 (2-3+A) = M-TH A-S (c) Characteristic Equation The equation |A- Al] = Ois called the characteristic equation of the matrix A e.g. 23-7272 + 114-5 =0 (@ Characteristic Roots or Eigenvalues A The roots of characteristic equation | A - A/| = 0 are called characteristic roots of matrix . eB. 23-7274+11A-5 =0 @-N)A-D)A-5) =0wpane 1 or or or or or eristic root S are 251 Hmportant Properties opr. (i) Any square matric 4 ee Mues Note. (2) The sum of the eigen diagonal of a matrix is called the trace (3) The product of the an t Bs fas Aa (2) WA Aare My are the eige (@) KA are kA kA Benvalues Hoesen Ky ‘alues, S of a ; = Ta matrix is equa nvalu qual to the trace of the matrix. S of ii) A“ are 1, Gi) AM are AAS, ae Example 59. c Examp Find the characterisn iomaere wa Sic roots of the matri. 6 -2 2 7 oo Solution. ara Tyee Se 6-2 -~2 a [ 7-2 3-A -1]20 2 ia (6-2) (9-6A4+22 A+22—1)42(-6 = —23 + 1222-36432 = foe oe 7 By trial, A = 2 is a root of this equation ‘A= 2) (2 - : ane (A=2) Q?- 1024416) = 0 or (A-2)(A-2)A-8) = 9 = 2,2,8 are the characteristic roots or Eigen values. Ans. Example 60. The matrix A is defined as 3 12-3 lo 3 2 00 -2 Find the eigenvalues of 3A} +5A?—6A+ 21. Solution. |A-AJ| = 0 1-2 2 -3 [ 0 3-Aa 2|=0 0 0 -2-A (1-2) (G—A)(=2-A) =O or AH 1H? Eigenvalues of A? = 1,27,- 8; Eigenvalues of A = 1,9.4 Eigenvalues of A = 1,3,-23 Eigenvalues of / = aad Eigenvalues of 3A? +5A?-6A +21 A First eigenvalue = 3” +5 (IP -6(1) +2 = 4 Second eigenvalue = 3 (27) +5(9)- 63) +21) = 10 Third eigenvalue = 3 (-8)+5(4)-6(- 242) = 10 Required eigenvalues are 4, 110, 10. Ans.pp 252 Matrices Example 61. 17 3 (Aap. Idi As Solution, (4-) 9? = 4 Ay are the eigen values of A, find the eigen values of the matrix RAIVEB = AP- DRAFT MoM Eigenvalues of 2A are 2A hy. 2A Ay 24g ww 2 A Eigenvalues of 02 /are w, Eigenvalues of A? are 3, 22 Eigenvalues of A?-224A+22/ are RADAR +I R= 2A AERA DA Ag tan “ Qa ARQ2= A Os AP one OA AS. os Example 62. Prove that the following matrices have the same characteristic equation abc) (bc a) (cab beal,|cabl,jabe cab] la bc} lbca Solution, Characteristic equation of first matrix is |A-A/| = 0 a-hk ob c or b c-A a c a bm) (@-A) 2-2 (+0) + be- a4] —b (BP -ac-bA)+c(ab-c+cA) = 0 or = +22 (atb+0) +A (-ab-ac-be+a? +b? +c?) + (abe - @ - b+ abe + abe- 0) = or A3-A2 (a+b +0)-A (a? +B? +c? - ab — be -ca) +(@+b+0-3abc) = 0. ‘The symmetry of the result shows that characteristic equation for the other two matrices will also be same. : Proved Example 63. Prove that a matrix A and its transpose A’ have the same characteristic | roots. ; . : i Solution. Characteristic equation of matrix A is HI |A-M = an i Characteristic equation of matrix A’ is ii |a-u] =0 7 know that } poth (1) and (2) are same, as we \ Clearly owe ] -. adeterminant remains unchanged when rows be changed into columns and columns into rows 1 ie, @ — i example 64. [fA and P be square matrices of the same type and if P be invenible, 1 -1 th oS i how that the matrices A and P~'AP have the same ae roots, sl , \MLLET.E., Summer 1998, Winter 1996) = P-!AP and we will show th: i . 1. Let us put B= PA that characteristic equations for ee are the same and hence they have the same characteristic roots, both A an’ BoM = Po!AP-Al = PAP-P-'NP = P-1(4_ ap p —_——act 253 IBM pPraeanpl = [poy |a- wile [Aa [pot pp = (Anal Let . JA~d| [1] = [A-adfas| A] =F s the matrices A and “ive - a reriatic 00s, Md H Tave the same characteristic equations and hence samme ple 65. IPA and B be 1 1B and BA 0 square ices, then prove that AB an vine charcteristte rape Me invertible matrices, then AB = IAB = B-'B (AB) = B>!(BA)B A) But by Ex. 64, matrices BA and B-! (BA) ; en : B have same characteristic roots 0 ‘pa and AB by (1) have same charac ee a Proved Example 7 ILA and B be n rowed square matrices and if A be invertible, show that the matrices A~! B and BA~! have the same characteristic roots. v1) Solution. A“'B = ABI = A~'B(A~! A) = A“ (BAW) A. But by Ex. 64, matrices BA~! and A~! (BA~!) A have same characteristic roots or matrices Av! and A~' B by (1) have same characteristic roots. proved Example 67. Show that 0 is a characteristic root of a matrix, if and only if, the matrix is singular. Solution, Characteristic equation of matrix A is given by |A-A| = 0. If A = 0 then from above it follows that |A| = 0 ie, Matrix A is singular. Again if Matrix A is singular i.e., |A| = 0 then JA-M] = O=|A|-Al/] = 0,0-2-1 = 0-92 = Example 68. Show that characteristic roots of a triangular matrix are just the diagonal elements of the matrix. Solution. Let us consider the triangular matrix. ay 0 0 0 Proved Aol eo 0 ayy a2 33 0 ay aan 3 as Characteristic equation is |A- Al ay-h 0 0 0 0 ay or =0 ay ay-2 0 an a3 ayy-d On expansion it gives (ay ~) (a22~ A) (aay =) (aug =A) = 0 * A = ayy ar, ayy ass “hich are diagonal elements of matrix A. Proved Example 69, The characteristic roots of a Hermitian matrix are all real. Solution, We know that matrix A is Hermitian if Ao = A ie, wherea® = (A’) or (A). -Matnices a 254 = BAY. ce NOAD = 2X Also (AA? = ZAP and (AB! IfA is a characteristic root of matrix A then AX (Ax? = AX)? OF But 4 is Hermitian XOX. ay ea) Again from (1) IOAN = XORX / } showing that 2 is real Hence from (2) and (3) we conclude that ne of real symmetric matr Deduction 1. From above we conclude that characteristic root > x are all real, as in this case, real symmetric matrix will be Herminan @) =A Rest as above. - zero or a pure imaginary For symmetric, we know that A’ = A. oe AS=A «A= Aas Ais real. 7 ither Deduction 2. Characteristic of a skew Hermitian matrix 1s ew number. If A is skew Hermitian then iA is Hermitian. Also A be a characteristic root of A then AX = *X- a G-A)X = (A). Above shows that i2 is eae ea matrix iA, which is Hermitian and hence i} should be real, which will be possible if 2 is either pure imaginary Or Zero. Example 70. The modulus of each characteristic root of @ unitary matrix 1s unity. Solution. If A be a unitary matrix, then we know A°A = 1. If A be characteristic root of A, then AX = 4X. _ (Ax? = (2X)? or XPA® = 2X® X°APAX = RXOAX = AXAXOX. But AA=IL. XIX = RAX®EX or XX ” (I-AA) XX = 0. But X # 0+. XX 2 Oandhence 1-24 i.e. characteristic roots are unimodular. . -) (2) A KX = Oorl = 2-Aorl[AP Example 71. If} is an eigenvalue of an orthogonal matrix, then 7 is also eigenvalue. (A.M.LE.T.E., Winter 1995) [Hint: AA’ = / if 2 is the eigen value of A, then 22 = 1, 2 = 24 a 4.26 CAYLEY-HAMILTON THEOREM Statement. Every square matrix satisfies its own characteristic equation. If JAA = (CD" tea, Nh ay 2s nxn matrix A= (aj), then the matrix equation ++ 4a,) be the characteristic polynomial of X" +a, X"-1 + ay Xm? . ” aX" 1 + arX"-7 +... + aq = 0 is satisfied by X=A ie,200 . 2 Shey ; Now thatthe main a= pL? | “ 1 ad] r ence find A evi canton Heme gt Verity c HY Cayley Hamitton Theorem for the mat 141 2 Py a | Wt Hence evaluate A ' a ; } -3 & 10, Use C " Une Cayley Hamilton ‘Theorem to find the inverse wre Te nse 7 re -1 5 2 Awl-10 3 a 2 M. HWA) Ay and Ay ate the eigenvalues of the mattis " " is equal 0 S28] ennge tet atem uae “9 A transformsvector X to vector Y. Two vectors X and Y have the same direction. Here we have to determine those vectors X whose images Y are given by Y =AX: Corresponding to each characteristic root % we have a corresponding non-zero vector ¥ which satisfies the equation |A-21|X = 0. The non-zero vector X is called characteristic vector or Eigenvectors.= Semmens el 261 naatnces jas PROPERTIES OF FI¢ 1. The cetor X 2 WA sectors Nis Xa VECTORS do be distinet ef Xe foun a tine 3. If wo or more » of am <1 A 4s not unique. orresponding eiBer cnvalues of an nxn matrix then © ey nearly Se dependent set , jet tint values ot be possible to B endent eigenvectors « are equal it may or may m6 uerend corresponding © the equal roots. oo Ne clucnvectors Xy and X; are called orthogonal vectors if Xy’ X2 = ee 5. Eigenvectors of a symmetric matrix corresponding to different eigen! orthogonal. Normalised form of vectors, py aa, 1 ws For example, normalised form of | 2 |is\ 24 2 * 429 NON SYMMETRIC MATRICES WITH NON REPEATED EIGEN VALUES Example 76. Find the eigenvalues and eigenvectors of the matrix , a vide each element To find normalised form «(tl we divide c i Oo 4 A=\1 2 1 22 7 Solution. |A-ar=0 1-2 0 “1 1 2-2 il-0 ie, B-62+114-6=0 2 20 3-4 LetA=1, 1-6+11-6=0 yi +1 By synthetic division T (A-1) 02-5246) =0 Q-1)@=-2)@-3) =0>4=1,2,3 To find eigenvectors for the corresponding eigenvalues we will consider the matrix equation (A-2ANX =0 1-a 0 -1 Yfx] [0 1 2-2 1. \lyl=I\o = ( 2 2 3-A\|z| |o Eigen vector corresponding to eigenvalue = 1 By putting A = 1, the matrix equation (1) will become 0 0 -1][x] [0 -2 11 iljy|=|O) => x+y+z=0, Letx=k 22 2Nz 0 k+y+0=0 y=“A y Solve by samer aube bythe eguation 5x4 BY,-4Z1- 4 =? Bo teBfe +b 22-6 HO “8% + bY, +2Z3-2=° preiae [f: -4 74 3 2 6 -6 SS 6 & 2 A) [eon 2 & 3 2 6 —2 bs 2 = 564-326) ~-30bt+4E)-A4CIF +16) = 5(-325°3154)~-4044). = -lbo-162-/36 | 458] 4,- [2 4 -4 2 6 -é& 6 2 2" 3 Clete Fy (-4 4 3b I~ 4 (4-36) = 3022+ 47225 ~4f-5m) Ar = 128 4 125 [A = 256)BO per wed ACH - stot al 7 Hale 216 a 23 Se 8 6 -2 2 50-4136 ose = 2004/26 -736 = 326-736 2 [bor [be —126 322-126 = Bb a = Gee, DH eee. Ae 54) bye) ~ Ul Or ary 454) pe 3lb- 48) 4G 9 3 0-54) 403) ee ee a. Ar As 4 2z- J 7 ; ee arb Ay 32 (ise 7 136 +b) = 492+ Ase BLY,2 Yale sub fn gn ©: Vy O-558) | 38-9432 24284 ol sb ASL we 458 ~ 1/822 = 2183 = 2.(8 76 = 2-183 22-1832 = 21: iy yy Solve by creer ate by the equation bxi+ BY, —27,-4=0 . 622 4442 +222, -6 =? ~2%3- 24, 97, p20 wee -~ [4 2 -2.-4 Wm oR Ss -«| “2-2 8 18 = it. a [4.42] -2 -2 = 60324-2048 +4) — 20-248) > 2/6 —/04+(-8) ae ee 4 2 -2 £ 18 = 20-36 +42) + 26-724+12)-4 (32-4) = 20 2)72(-60) 4628) = 24-/20-/12 2s 24-222 - 28 ae Ca a7 6 2 -& -2 @ -&L (- 36448) +2 (-108-I2)- 4 (4844) 6 (12d t 2C- 1209-4 Cs2 " Jt - 24 0-28 " A yee ea | 4:-f4 2 74 O42 | ——- = _& . BC HTR HA) 2 OH 108 = AD =4 278) ee ee ~3b0ere4ye- FO = Gyo F240 202 4224, A= -208 , As=-376, As =-2080 se 2 LZ LH! 4, =e Py ES = “J SZ a! hi 208 —376 too 7 20 — a aS 22 208 29. 6666 m20F Ga! 24, oF -) 7 = Ob DS ye =37e - 15-6666 376 aad 2 24 2 Zz -1 =. ss z= —202 . &.33323 —-oo “ay 3 a y 24 xyz valet Subih e9n G- 86666 15.6666 _ 9.3333 — 208 376 -2eo ae —O8-04)] = -0-04;, => - 0-041 = =e calcule by the Guat, \ BZ Solve ay Cramet age Z=! Bx tbyp bE =4 Fxtay-32° Ca We seryyz-1 0 B24 5ypb2-4 =P Terey 2" ° Ao = ie) Ss 6 _4 2 - =) = : 2 a p 7 2- lee) 21 (-9- F441 CE 45) = ~277+6 3-39 1-27 ~1(-63) 410-37) 3 ie ed P44] S 1 (-102- 12) - CPS 48-10“ 12) = 1-114 —C- 77D ~ 12 7) IAF TV I FRT = — /AA4 +loy t ,o-? S 6-4 <3 7 = (l-lo2-f2)-10-5 14 36)-1C- 9-54) 10-114 )~ 10-16) - 10-63) " \ r " MI 415 4L32 1-854 8-1-5 BEI 16-459 M-77) -10- 5) - 10-39) " ae 774 ISH 39 = — 77454 2 2-35 515-10, Ad tas ao 2 ao —3Be 23 De As ders i —lo 3 2 L = 3& = 12 ae AO Be 3 . =22 2 7. 6666 Zz =) 2 = Be 3 -23 ‘ 73 3333 y2 AT-GEEE 723, ©-3323 = 0-3333 - 2.3333=°:-33 vracteristic equation of matr;, « of order”. Suppose there jg a," that the action of A on x is just a multiple of X, ina» Is a1 Fiset valu OT fet A= a} jimetional onze CO fhe matrix product AX, &! ne ma Nn ‘ er words, the transformation represent, where Ais a seat vector X by ascalar A. The vector X a by ora. dis called an Eigen value of tor X . The problem of finding the eigenvect, tor called the eigenvalue problem. ts igenva’ a : and oe faition of Eigenvector. Anonzero vector x - called an eigenvector ofa matrix A if there isa number A such that AX = i Here Ais called an eigen value of A corresponding to the eigenvector x and vice versa. We have, AX =AX =AIX, I being a unit matrix. ce (A-ADX=0. This is the matrix form of an eigenvalue problem. Since X #0, the matrix ( A-—AI) is singular, so that 1A-AII=0 (1) Equation (1) is called the characteristic equation of A The eigenvalues are just the roots of the equation obtained by expanding the determinant in Eq. (1). The n- roots Ay, Ags . A, of the characteristic equation are not necessarily all different. Example 1. Determine the eigen values and eigen vectors of the matrix: A=/[3.14 Sol. The characteristic.equation of A is | A — AI | =O i.e. 3-A 1 4 ° 0 2-A 6 0 0 5-2 These are the eigen values of A . To determine ei igen vectors let us F @ When A, consider oe the eigen values one by one. e eigen vector X; is given by (A — 21) X, =0 1301 paice : sale rank of coefficient matrix being, 2, the equation will have onl: | eT inearly independent solution ' 3-7 Tose are equivalent to x) 4.43 4 4xy=0 phe last two give x: . Then first one gives x1 + x9 =0. ake x1 = 1, then x) =— 1 and x3=0. Hence Xyscy ‘ C, being a scalar. o (ii When Ay =3, the eigen vector Xj is given by (A — 31) Xj =0 O11 404 0 ies 0 -1 6|\\x2\=|0 0 0 2\\x,| |o These are equivalent to x) +4x3=0 =x) + 6x3=0 2x3=0 giving x3 = 0, x. =O and x is arbitrary, say x, = 1. &,=C,[1 Cy being a scalar. ‘Then 0 : 0 (iii When A, =5, the eigen vector X3 is given by (A — 51) K,=0 —2 1 41/41) fo ie., 0 3 6\|x\=|0 0 0 O})x,) {0 These are equivalent to —2x, + x) + 4x3=0 3x, + 6x3 =0 giving X= 2x3=3x,, ie, 2x, =3x)= Ory Take x3 = 1, so that xy =2 and x; =3 Les %=[s C, being a scalar. 1> Mechani 4 values am . he eigen ons Or” matris io teristic uation of A is The characteristic eq a ° ie 1A-All=]| 0 ei 0 (1-a) [1 -4)?- 1] =0 (1-A) @?- 2a) =0 Le, he, hen 2(-a)(A-2)=0 te 2=0,1,2. Thus the eigen values of the matrix A are 0, 1, 2. Eigen value equation is (A-AD X=0 For 1.=0, Eq. (1) reduces to + ca 0 o--F [1]. fo . o11 x2/=/0 On 7 ay 10 This is equivalent to the following equations . %=0 42 +23=0 : %2+x3=0 Solving these equations, we get 41 =0,x) =~ x3 = (say) *) = X= X2]= fx, 2b) 14) = iy 1233} ={ 0,4, -k} a normalised eiZen vectors ex and Mathematical nqq,, he dy th, -- (1)be 133 Matrices ifthe eigen vectors be normalised to unity, then 1X4 t= 1 1 of Vf02+ 2+ (CW) = tork= or +(—k)y) =1ork =I fe Normalised eigen vector _ 1 -1 oo {o-xtay Te} For A= 1, Eq. (1) reduces to 0 0 oj("| fo 0 0 1\\2\=\|0 0 1 0\\x,| |o This is equivalent to the following equations x, =0 x3=0 =. X= {1, 0,0} in normalised form. For 4=2, Eq. (1) reduces to 1.0 0|\*t 0} o -1 1\\x2|=|0 0 1 -I\\x, o| which is equivalent to the following equations =x, =0 -x2 +x3=0 x -x3=0 Solving these equations, we get x1 = 0,2) =%3 en vector corresponding to Within the arbitrary scale factor, the eig 2 =2is given by X3= (x1 %2,%3) = (0K KY For | X31 normalised to unity : L +e+Rel ie, kay X3= {0 oe a in normalised form. Thus the normalised eigen vectors of the given matrix A correspond- ing to the eigen values 0 , 1,2 are fo. 7: -3} , (1,0, 0} fo, aoa + respectivelyMechanics and Mathema, Hicay : M 134 8.2. Cayley-Hamilton Theorem 7 Statement. Every square matrix satisfies its own characte,;, ‘ OR cea, If Ja —al] =a9 +41 A442? +... +a,A" = 0 oy be the characteristic equation of a square matrix A, then ag +a, A+a2 A? +... +4, A"=0 Proof. The characteristic polynomial is 1A —AD | =ap + ajA+ ay? + 4,4" Each element of characteristic matrix (A — AI) is an ordinary po} ~(1) of degree n (at most). Therefore the co-factor of every element of A YAOmiay ynomial of degree 1 - 1 (at most). Consequently each aul is lemeny an ordinary pol: . of B=adj(A-AD) is an ordinary polynomial of degree (n — 1) (at most). ~Q B= adj (A- Al) = Bo + ByA + ByA? + + By_y AP—! c ~@) Here Bo, By, Bz,.--By_y are all square matrices of the same order whose élements are polynomials in the elements of A. . “Now, (A — AD) adj (A-AD=1A—All 1 Using Eqs. (3) and (1), we get (A-AD [Bot By A+ Bz 47+... +By_ 127] =(agt+a,h+azA?+... +4, AT 4) Comparing the coefficients of like powers of 2 on both the sides, we get ABy = al AB, - By =a, AB, - By = a1 AB,_1—B,_2 —-B,_)}=4,1 Premultiplying these by I, A, A2, A> ---- A in order and adding, QI + a,A +a,A7+4---+a,A7=0 a,_11 This is Cayley-Hamilton theorem. Corollary. To determine A~ 1 by using Cayley-Hamilton theorem. Let A be a non-singular matrix of order n so that 1A! #0.Matrices According to Cayley-Hamilton theorem 135 : gh + yA + GAZ 4g Ane ‘The characteristic polynomial is nA" =0 Ay (AWM sag + ay Aah? + gan ? . Bq. (2) gives |Al=ay. - ao eo (2) é ay iding Eq. (1) by ao, we get t=-|O 44 2Q2, 2 a9 a (3) Pre-multiplying Eq. (3) by A~!, we get Ala -|%y,2 a, we ag ong pn dq tt ay At tEAM 1 £4) Example 1, Find the characteristic equation of the matrix 1 23 A=!2 -1 4 3 11 and verify the Cayley-Hamilton theorem for it. Hence or otherwise find A". 4 123 100 3 Sol. A-AI=|2 -1 4]/-A]0 1 6] = 4 3071 oo} {3 1 o1-A 3 4 |= -23422+4181+30 3 1 I-A’ Hence the characteristic equation is . 23-422 + 182430=0 Now, in order to verity Cayley-Hamilton theorem, we have to show that —A3+A?2+18A+430F=0 100 123 HereI=|0 1 O}, A=|2 -1 4 oo1 3 14 1 2 3]/f1 2°3] [143 14 A2=/2 -1 4]|2 -1 4/=|129 6 Bet iio. tot 86u 143 14][1 2 3] [62 39 68 A3=A7A5|12 9 .6||2 -1 4] =f48 21 78 8 6 14/{3 11 62 24 62 2. ~A3 + A724 18A+ 3anics and Mathemancal Mer, tho, ‘ds ‘ Meche ™ a a la Lot2e3 100 o2 3 . 2 9 of +18] -1 4 ae 10 6 BG 6 a i of > 24 6? 62 ? 9 00 900 =0 000 — Cay! id - Hamilton theorem IS verifies Hence ©4 _ To find A~ : 301-A° 34A74+18A J ; 7 1= 39 (47-4 18 A) 1 we get ying the above equation by A~ Premultipl -12h (a?-A- 18D A 30 ¢ 3 14 Te eA ag 9 6] -b}2 -1 4/-30/° 3 ? 6 14| 7{3 1 tt oo1 5 1 wu -30 30 30 2 1] = = tees a 30 30 30 Example 2. /lustrate the Cayley-Hamilton theorem for the matrix 120 A={2 -1 0 ool 1-A 2 oO 1A-ALI=] 2 -1-A 0O oO 0 1-4 = -5+5A4+)7- Hence the characteristic equation is , -545A+27-A3=0 We have to show that .— 5I+5A +A?—A?=0 Now, -5I+5A+A?- 0 -5 0 0 5 10 0) (500 5 = -5 -1 =| 0 -5 o/+/10 -5 0/+/0 5 O0]+}-10 ao 0°40 5 0 05] joo1 Oe 0 1matrices 8.3. 4 137 gxample 3. Find the characteristic equation of the matrix 2-101 A=|-1 2 -1 1-1 2 satisfied by A. Hence find the inverse of K « fy that it and veri sol. The characteristic equation of A is 2--’ =I 1 IA-AIl=] -1 2-4 -1]=0 1 -1 2-2 Les —23 + 642-9244=0 23-607 +92-4=0 (1) ie, ‘We have to show that A3—6A24+9A -41=0 5 5 22 -21 21 6-5), Ar=|-21 22 -21 A UAa =o 0 21-21 22 - A3-6A7+9A —41 22 =21 °21 6-5 5 2-107 =|-21. 22 -21]-6|-5 6 -5|+9/-1 2 -1|-4 1-1 2 00 . 0 21 -21 «22 5-5 6 ol oo- A3-6A74+9A-41=0 (A3 — 6A? +9A) Tofind A~*: Premultiplying the aiove equation by A~ A71=4(A?-6A+91) 1 we get 6-5 5] fez, 6. -§] ,J209° a1)-5 6 -5|+4| 6 -12 6|++|0 9 0 Sits 5-6 67°. 6%=12 009 fat ot =4}13 1 me Theorems on eigen values and eigen vectors Theorem 1.The eigen values of a Hermitian mal Broof. For a hermitian matrix A AT=A Here, A ¢ is the transposed conjugate of A - trix are all real.i Jb Mechanics and Mathematical jy eu 138 _ Let A be an eigen value of a Hermitian matrix A. Then a8 here exists a vector X #0, such that hin : AX=AX. 4) Premultiplying Eq. (1) by X t, we get Xt AX=Xf AK =AXtX (2) “Taking transposed conjugate of both sides in (2), we get (Xt AX)t =@X4X)t Xt AT KNT=A+ XTKDT pas XTATX =A*X1X or X (AK =2*X7X (At=ay XfAK=A*XtX (sing Ey, 1 or (-A+)X#X=0 8) As X is an eigen vector X#0; -. X{X+0 Then Eq. (3) gives, 2-2 =0 or A=2* This means that the conjugate of A is equal to itself. This is only possible when 2 is real. ‘Thus the eigen-values of a Hermitian matrix are all real. eorem 2. The eigen values of a real symmetric matrix are all real. Proof. For real symmetric matrix A, A*=A and AT=A S (AT =A ie, ATH=A ‘Thus real symmetric matrix is a hermitian matrix. By Th. |, the eigen values of xhermitian matrix are all real. Consequently, the eigen values of a real symmetric miatrix are all real. eorem 3. The eigen values of skew-hermitian matrix are either zero or po imaginary. ‘Proof. For a skew-hermitian matrix A, Aj=—-A (1) Let X be an eigen vector of A corresponding to eigen value A. Then AX=AX ---(2) or ~ GA) X=(@)x --@) Now, GA)t=#At=~i(-A) using (1) * (iA)T=iA This proves that iA is Hermitian matrix. According to Eq- (3), iA. is the eigen value of hermitian matrix iA corresponding to the eigen vector X. Therefore, by theorem 1, 21 is areal number. It follows that is either zero or purely imaginary number.Se} ! SC [2 19 psriceS ’ ue 4.The eigen values of a real skew-symmetric matrix are either orem . weer magna: ° vor a real skeW- symmetric matrix A, rool. F et ee At = —A = Ais skew-Hermitian, 2 py Th. 3 the result follows. f . The modulus of each eigen value of a unitary matrix is unity eae Meet esa, fier A) yet X be an eigen vector of A comesponding to eigen value 2. The, ren (2) ‘Taking transposed conjugate of (2), we get (ANT =AX)t 2 XTAt=AeXt - Post-multiplying (3) by (2), we get (KTAN (AX) =(* KH-AX) X#(ATA) X=A0 + X4X or or XTX =20XHK using (1) or X#X =AA*XTX or X#X(1-2A7)=0 4) As X#0, XiX#0 Eq. (4) gives, 1-2A"=0 ie, AAT=1¥71=1 ‘ IM=1 ie, the modulus of Ais unity. Toenrem 6. The eigen values of an orthogonal matrix are unimodular (ie, of unit modulus) Proof. For an orthogonal real matrix A, we have A‘=A (1) ATA=I (2) ‘Taking complex conjugate of Eq. (2), we get (ATA)* =I" or Grrr £ rs = ATAS=I using (1) unimadg(® i unitary. By Th. 5, the eigen values of a unitary matrix are unimodular Hence the eigen values of a real orthogonal matrix are n % vale BEOENLZ. TELL Any two eigen vectors corresponding to two distinct eigen ‘ermitian matrix are orthogonal.a bien lL , Mechanics and Mathematica M, eth, » 5. X, be two eigen vectors correspond os A a en ae AThen eo distin eigen values Ay. Az pa AX, =A,X, Gi AX) =A)X2 ~Q) From Th. 1, Ay and 2g are real numbers. 3) Aish haa 3a” (4) Premultiplying (2) and (3) by Xpf and X,+ respectively, we pet Xpt AX, =A,X9FX, (5) X1FAX2 =X) 7X) Taking transposed conjugate of (5), we get aS) (XQtAX))t = A,XQtX pt XAT DT=A, Xt Kt ; Xyt ATX, =A) XpFXy Using (4) ce X FAX = AX 7Xy using (1) ..(7) Comparing (6) and (7), we get AX 7Xq = 9X 7Xy ie, (Ay 2g) X#X_=0 Since A, — Ag #0, otherwise the roots will not be distinct, the only possibility is that X;+X,=0. It follows that X,, X, are orthogonal. Theorem 8. Any two eigen vectors corresponding to two distinct eigen values ofa real symmetric matrix are orthogonal. Proof. Let A be a real symmetric matrix. Then A‘SA AT=A Taking transposed conjugate of (2), we get ce (AT)*=A* or ATSA using (1) =A is Hermitian. By Th. 7, the result follows, Theo ny two eigen vectors corresponding to two distinct eigen values of a unitary matrix are orthogonal. Proof. Let A be a unitary matrix. Then AtA=I wD)141 «pe two eigen vectors corresponding (0 Ava distinct eigen i Mra of unitary mates A. Then o Kp = aXy 2) AX) =2Xz G3) cposed conjugate of (2), we get aking transpose . (AX))t = uXpt or X PAT = Ay "Nit Ad) post-multiplying (4) by (3), we get (XjPAD (AX) = Ah “X17 A2Xo) 7 Xt (AFAY%) = Ay “AaX1P% or Xt1Ky = Ay "AoX1 7X2 using (1) or XyfX_= Ay “AX FXo ( X2=IX) or (1g "Ag) Xi tX2=0 (5) rix, the modulus of each of its eigen values But A being a unitary mau is unity, be. Ayth=l (1A MQ) =O“ *Ag) = Ay “Ay #0 (since Ay # Ay) (6) From (5) and (6) it follows that X)+X2=0 ie, X, and X,are orthogonal. Eigen values of a matrix are invariant under a similarity Theorem 10. transformation. Proof. Considt transformation such that ler two similar matrices A and B, related through similarity B=P-'AP ‘The characteristic equation of the matrix B is |B—M1= 0, Here, 2 is an eigen value of B. ie, 1p-!AP-AI=0 or ip! ap-P-! AP 1=0 | or 1p-'(a-anP or IP7'1A-MIIPI=0 oe ipo! Pi A-M1=0 IA-Al1=0 Thi * . his relation shows that 2 is an eigen value of A also. This proves the theorem.met a Mechanics and Mathematieay if , en Meth, corem HL, The eigen values of a diag onal matrix are prec, ; ely 4 clements in the diagonal. re Proof. Let A= diag. ayy, ay), °° Annl Then (A ~ Al) =d [ayy A, ay) —A, gy Ad ‘The characteristic equation | A —AL1=0 gives (ayy A) (497-4) (an 2) = 0 hea 4, 027, --- Ging 11-422" Gyn a6 the diagonal elements of A. Therefore the cige ; . A n values 2 of a diagonal matrix are the elements in the diagonal Dipgonalization of matrices. Let Ay)Ap"- Ay be n distinct eigen values of a XX matrix A ang -X,, be the n corresponding eigen vectors, Let X; be the column vector given by xy Xj=| Xai (I) Xni Consider a matrix P whose column vectors are n eigen vectors such that Yi Xp p=| %21 Xx 2) mt Xz... --Q) MX AX, Then PD =| 24X91 4Xoy (4) Xm Xap (n0 summation over j ) = (AX), AX) ... AX,) (expressing matrix as vectors) =A (KX)... X,) =AP..f- matrices 143 -singular matri: iP be a non-singular matrix, then premultiplying (5) by P=! } D=P-' ap We get | tiplying A by P~! ane 6) | Thus, premultiplying A by P”! and postmultiplying by P, we ger the hus Pre whose diagonal elements are shee . waz diagonalization ofthe matrix A. igen values. This process is ca aia al example. Let A=|4 _3 ‘The characteristic equation is is 1A-All ie -3a)-° Le -9+22-16=0.0rk=45 ies Ay=-S and 2)=5. ‘The corresponding eigenvectors are easily obtained : -1 x= =| and »-[j| _[-i 2 Lat rs i -1_.1) 1 2 Then P 3|-2 -1 -l apa — 2) |3 4) |-1 42 Hence, P~ AP = | F 3 a | ! ule ape ule u ° — Pook b- b- Ld wy 1 Su ud, z ° Practical Method of Diagonalization. To reduce a given square matrix A to diagonal form, we first write the characteristic equation for the matrix and evaluate the characteristic roots 24, X --- Ay- Then the required diagonal matrix D of A will be m0 0 1.0 : 0 %o .. 0 D=|0 0. 4... 0 000 Mn . Example 1. Diagonalise the following matrix cos@ —sin® 0 A=|sin@ cos® 0 0 olMechanics and Mathematical Methou ls i cos ~sin@ 0 A=|sin@ cos@ 0 sol. Let 0 ol e cteristic equation is : nee cosO-2 -sind 0 5 jl 6-2 0 |=0 \A-Ml= i iG han (1-2) (1-2A.cos +27) =0 2cos 6+ V4 cos” 0-4 2 ie. Characteristic roots are 1, 1, cos 8 #isin® 1,8 : a 24 = 68,2. =e" ®, 25 =1 (say) - The required diagonal matrix is ie, ie, 2% 0 0 D=|0 e-®0 0 01 8.5 Matrices in Physics (@ The Rotation Matrix. Let (x, y, z) bea cartesian coordinate system ina three-dimensional space. Let u = { y,u9} be a vector in the xy-plane. Consider a rotation of the coordinate system about the z- axis through an angle 6 in the anticlockwise sense. Let us denote the new Coordinate axes by xyz. If the same vector u has components u,’, 4’ relative to the new system, then 4)’ =u) cos @ + uy sin @ 4! = uy sin + uy cos 8 --(1) This could be written in the matrix notation as “l_[ cos@ sing uy 2'}"L-sin@ cos} |u, If we denote the rotation operator by R, (8), we could write the transformation of the vector w to symbolically by Usu=R@)u ~@) The operator R, (0) which causes “ector u tow’ is represented by two dr the transformation of two dimensional mensional matrix R -[ cos@ sing 24) ~sin® cos @ andis called two, dimensional rotation matrix. It is an orthogonal matrix.