Solution. The characteristic equation of A is a/2-v 2 . V2 3/2-a| 70 be, @R-AP— 4 =0. 22—84+2=0 whence A= 1,2. When d= 1,(A-MIX=0, gives (ie wall]-L2) » GF lle )-Le (2k 2 " , [o a [2]-[0] [By R,-R,) x,+x,=0.Ifx,=—1,x,=1, ie, the eigen vector is [1,—1]'. When 2=2,{A—MIX=0, gives Fu alle )-(4] i (3 -ill2}-Le) [i] . faba enn =x,#x,=0, ie, x=% x,=1, ie., the eigen vector is [1, 1] Replacing e by 4, we got 1/20 12]_flo 6 12 20|"| 6 10]° Ba REDUCTION OF QUADRATIC FORM TO CANONICAL FORM A homogeneous expression of the second degree in any number of variables is called a quadratic form. ah x For instance, if hob ‘|. y| andX’=[x y 2], then Bf ¢. £ KVAX = ax? + by? + 2? + 2fye + Qgax + Qhxy oi) which is a quadratic form.| Livean ALosibna : Derentnawrs, MaTrices Let Ay, Ayy Ay be the eigen values of the matrix A and be its corresponding eigen vectors in the normalized form (i.e., each element is divided by square root of sum of the squares of all the three elements in the eigen vector). a 0 0 he % % ‘Then by $2.16), P*AP=|0 2, 0|whereP=|¥ 92 33 0 0 dy 4% 4% Hence the quadratie form (i) is reduced to a canonical form (or sum of squares form or Principal axes form). Tyr? + Agy? + Age? and P is the matrix of transformation which is an orthogonal matrix. Note. Congruent (or orthogonal) transformation. The diagonal matrix D and the matrix A are called congruent matrices and the above method of reduction is called congruent (or orthogonal) transformation. Remember that the matrix A corresponding to the quadratic form ax? + by? + c2* + 2fye + Qgex + hry coeff. of x” pool of yz } coeff. of zx whe is J coett of yz coeff of »? Leveft.of xy|, & i 6 ‘| 2 2 efe 1 1 2 Meoott of 2x evel of xy coef of 2 poaoicedaeys ae teeth ha dit —Byz + Bex - joel 3 Rare 20) 3-1 1 Solution, The matrix of the given quadratic form is A = [- 1 5 - | 1-1 3 B-A -1 001 Its characteristic equation is |A-M|=0, ie, | -1 5-A -1 1-1 3-2 which gives A = 2, 3, 6 as its eigen values. Hence the given quadratic form reduces to the canonical form Aart Age+ag®, ke, 2a? + By? + Ge? To find the matrix of transformation From [A - A] X = 0, we obtain the equations (=A) x-y+2=0;-x4+6-Dy- Now corresponding to A = 2, we getx—y +2 =0;x-y+(3-A)z=0. x4+3y-z=0,andx-y+z2=0, shenee wens tO 1 ‘The eigen vector is X, (1, 0, 1) and its normalised form is (1/ V2 , 0, - V2). Similarly, corresponding to 2 = 3, the eigen vector is X, (1, 1, 1) and its normalised form is (1/3 , W/ V3 , vB). Finally, corresponding to 2. = 6, the eigen vector is X, (1, — 2, 1) and its normalised form is (1/ 6 ,- 2/6 , wJ6). [ wv2 wVS WE Hence the matrix of transformation is P= 0 W/V3 -2/V6|. <2 VWV3 V6ca -—Hiaien Enanserma ManiEnarics NATURE OF A QUADRATIC FORM Let Q = X" AX be a quadratic form in n variables x), 95 --%q- Index. The number of positive terms in its canonical form is called the index of the quadratic form. Signature (S) of the quadratic form is the difference of positive and negative terms in the canonical form. Ifthe rank of the matrix A is r and the signature of the quadratic form Q is 8, then the quadratic form is said to be ( positive definite ifr =n and s =n (ii) negative definite ifr =n and s = 0 (iii) positive semidefinite ifr 0. (ii) negative definite if all the eigen values of A <0. (iit) positive semidefinite if all the eigen values of A 20 and at least one eigen value (iv) negative semidefinite if all the eigen values of A $0 and at least one eigen valu (v) indefinite if some of the eigen values of A are positive and others negative. SR Rai tt adr fon Ba 3 eer aeey discuss its nature, i Va A pe ssednde As h 0 Solution. (i) The matrix of the given quadratic form is A = | 1 1 -A 1 1 Its characteristic equation is [A—2]=0, ie, 1-a- - i =0 1-1-2 which gives 4932+ 2=0 Solving, we get A= 1, 1,-2as the eigen values. Hence the given quadratic form reduces to the canonical form Aye + hy? +hg?=0, ie, 22 +y?-2e2=0 (i) Since some of the eigen values of A are positive and others are negative, the given quadratic form is Indefinite. (iii) To find the matrix of transformation From [A — XI] X= 0, we get the equations met y+2=0,x-hytz=0,x-y—he When 2=-2,weget 2x+y+z=0,x+2y—z=0,x-y+22=0. Solving first and second equations, we get ‘The corresponding eigen vector X, = (— 1, 1, 1) and its normalised form is (~ 1/V3, 1/ V3, 1/V3) Whend=1,weget —x+y+2=0,x-y-2=0,x-y—z=0. These equations are same. Take y = 0 so that x = 2. ‘The corresponding eigen vector X, = (1, 0, 1) and its normalised form is (1/ V2 , 0, 1/ V2) To find the eigen vector X,, = (I, m,n) (say) Since X, is orthogonal to X,, —l+m+n=0 Since X, is orthogonal toX,, -. l+n=0 L ‘These equations give 1 ‘The eigen vector X, = (1, 2, ~ 1) and normalised form is (1/V6, 2/V6,—_1/V6). 2Lingan Avceena : Deverwnants, Marrices Hence the modal matrix is -W3 W2 WE W302 |. W3 WV2 -VV6 recess =1/2 - 3/2 0 2 V3/2 0 |. HA=|-J8/2" 1/2 0] and P=|-Y3/2 1/2 0}, show that P* AP is a dingonal matrix. 0 00 o oL 2. Show that the linear transformation Hee fin 6]. where e= 3 tant 24 , changes the matrix «fe 5 to the diagonal form D = HCH". 3. Reduce the matrix A = ste x 1 tothe aia form. (BP.TWU, 2005) 4 was[f §] find at and at (Mumbai, 2006) a-1. t 6, WA=|-1 5 -1), calculate A‘, (Coimbatore, 2001) 1-2 8. 6. was[} {]s then prove chat 9 tan A= A tan (Stumbai, 2006) 6-2) 2) 7. Find the eigen vectors of the matrix |—2 _ 3. —1] and hence reduce 6x? + 3)? + 3z*~2yz + 42x — day to a'sum of 2-1 8 squares’. Also write the nature of the matrix (Calicut, 2005) ‘8. Reduce the quadratic form 2xy + 2yz + 2zx into canonical form. 10. (Anna, 2009 ; Kurukshetra, 2006 ; Muribai, 2003) 8 Sanh (6) Reduce the quadratic form x,? + $x,? + 8x32 —2e,x, to.a canonical form, (Anna, 2009) Reduce the following quadratic forms into a ‘sum of squares’ by an orthogonal transformation and give the matrix of transformation. Also state the nature of each of these. (8) 9e,2 4 S432 + Sry! + Beye + 2 yx4~ Dees. 1 0 Oo (i gat i ter oe ed ot 3 i} (ii) Bx? + Ty24 Be? — Ly — Bye + dx nna, 2002 8) 11. Find the index and signature of the quadratic form xf + 2x3 - 3x3. (Madras, 2006) 12, Find the nature of the quadratic form x* + 5y? + 2” + 2xy + 2yz + Gex. (Bhopal, 2099) 13. ‘Show that the form 5x,?-+ 26x,2-+ 10x, + dx, + Mdxge, + Gxt» is a positive semi-definite and find a non-zero set of values of x,,9,3 which make the form zero. COMPLEX MATRICES (P.T.U., 2003) So far, we have considered matrices whose elements were real numbers. The elements of a matrix can, however, be complex numbers also. (1) Conjugate of a matrix. If the elements of a matrix A= a,,] are complex numbers 0, + i Br. 0, and By, being real, then the matrix