Determinants and Matrices PART-A: Summary of Important Concepts A matrix is a rectangular array of numbers, arranged in rows and columns. For example 217 M= isa 2x3 matrix -13 0 Matrices with the same number of rows and columns are called square matrices. The determinant is a value associated with a square matrix, ie., every square matrix will have a corresponding determinant, which can be computed according to a specific arithmetic expression, For example, the determinant of a, 4, a la, a, a, M=|6, b, 6, \canbewrittenasA=|b, b, 6, So Sy. 2 Ss) and will be given by A =a, (b,c, ~byc,)+ a, (bye, — ye) + a3 (Bye, — bc) Matrices and determinants find application in an extremely wide variety of scientific fields, and this makes them indispensable tools for science students. Since determinants are quantities associated with a special type of matrices (square matrices), we summarize their properties first, and then tum to a discussion on matrices in general. 1. Expansion of Determinants la 5 A2 x 2 determinant A=|" "| has the value A=ad—be. To expand a general determinant le (of size n x n) along any of its rows or columns, we need to understand the concept of co-factor. If a, is the element in the ith row and jth column, then ignoring the ith row and jth column in the n x n determinant leaves us with an (n—1) x (n—1) determinant (xn determinant) M,((n— 1) x(n 1) determinant) i Ignore the ith row throw, ‘and jth column sth column, If the reduced matrix is donated by M,, then C,, the co-factor of a, is given by C, =(-1)"/M,.252 Determinants and Matrices Now, suppose that we denote the rows in an m x n determinant A by R,, R,,..- R,, and the columns by Corns C,. To expand A along the pth row R,, we find the co-factors of the terms in R,, If the terms are d, .d,,,.d,,,we denote the corresponding co-factors by C,, Cp, yu C,,. The value of A is now given by ° a= 4, C+ 4p,Cp,* 7 +4,.C,, That is, take each term of the row, multiply it by its co-factor, and add all such terms, Similarly, A can be evaluated by expansion along any column, It is important to note that no matter what row or column. you choose to expand A along, its value will always come out to be the same. As a concrete example, lets take 132 0 75 -2 and evaluate it by expansion along the 2nd column, We summarize the calculations in a table: Position of Term Term Co-factor Term x Co-factor a2 3 B @2 0 0 (.2) s 30 A=48 ‘You can verify that the value of A will come out to be 48, no matter what other column (or row) we choose to expand 4 along. Let us find the co-factors in a 4 4 determinant. 13 2 -I Ilustration 1: Let D=|° 2 4 7] Find the co-factors of 6, 7, and & 413 9 is 5 2 6 Working: 6 is at the position (4, 4) so i+ j =8. The co-factor of 6 becomes 132 0 2 4|-i|? 4 o4 o2 2-3x(-16)+2x(-8) = 3: S = 1x2-3x(1 x (-8) = Ana 13 a3]"laa 19+ 4 is at the position (2, 4), so i+ j=6, The co-factor of 7 is la 2 3 s qf =te¢c17)-340-20)+2402)=103 oS 3 1 5PART-A: ummary of Important Concepts 253 For 8, i+ j =5, s0 its co-factor is 32-1 C)]2 4 7]=-65 (verify) 139 ‘We note the following important facts: (a) The sum of the products of the elements of any row or column with their corresponding co-factors is equal to the value of the determinant. (b) The sum of the products of the elements of any row or column with the co-factors of the corresponding elements of any other row or column is zero. Properties of Determinants To evaluate a determinant, it isnot always necessary to fully expand it. There are a number of propert of determinants, particularly row and column transformations, that can simplify the evaluation of any determinant considerably. These properties are summarized below: Property 1 The value of determinant is not changed when rows are changed into columns and columns into rows. 4 a, 4| Ja & For example, A=|5, &|=]a, b oa Go| la & 6 Property 2 If any two rows or columns of a determinant are interchanged, the sign of the determinant changes but its magnitude remains the same. a 4 4% a a 4 For example, A=|5 b &)=-lq a 6 4% bb by Property 3 A determinant having two rows or two columns identical has the value zero. For example, r r Zz Property 4 Multiplying all the elements of a row (or column) by a scalar (a real number) is equivalent to multiplying the determinant by that scalar. For example, ha, ha, hay a4 4, 4 A=| 5 6 6 l=ala mb 6254 Determinants and Matrices Note carefully that 4 is multiplied with elements of just one row and not of the entire determinant. Another example ha, Aa; hay a a, a Ab, Ab, Ab, |=2? |b, by by Re, Rey Aes Ga 6 Property 5 A determinant can be split into a sum of two determinants along any row or column. For example, a+d, a,+d, a,+d, a a a,| |d, d, d, 4, b, b |= |6 o& b|+] 4% 4a 6] laa 6 ‘The property is evident by expanding the determinant on the LHS along R, Property 6 Row and column transformations. The transformations property is the most widely used property to simplify determinants. This property say's that the value of a determinant does not change when any row (or column) is multiplied by a scalar (a real number) and is then added to or subtracted from any other row (or column). For example, lets consider 4 a 4 A=|5, bb oa 4 Letsmultiply the 3rd ow by 2. and add to R,, This operation can be succinctly denoted as R, > R, +ARy. Our property tells us that the determinant’s value stays the same. Indeed. ate, a,+he, a, +e 4, by b RR +AC, “ o 6 a, a, a,| |ac, acy dey =|5 6 &|+]b b | — Gplitting along R,) oa al la a@ «6 a4 a 4 6 al la a a =|6, ob) +2) & B=] & | =a Go I Go & In the second-last step, we have used the fact that since two of the rows of the second determinant are the same, its value is 0. Property 7 Multiplication of determinants. Suppose we have two 2 * 2 determinantsPART-A: Summary of Important Concepts 255 and we wish to find A,A,. aaje| 2 [[% & a b,\|a, B, _|4 &|/@, @,| (We interchanged the rows and the ~ |a, 6, || B, B,| (columns for the second determinant. tontye(®) conto>{ ; touto(®) centr{f) In the last expression, R, and R, denote the rows of the first determinant, while C, and C, denote the columns of the second determinant, This is a row-by-column multiplication. Similarly, we'll have column-by-column multiplication, What if we have to multiply two 3 x 3 determinants row-by-row? aa, +ba, a,B, +b; 4,0, +b,0, af, +b,B, RC RC, RC Ry a & 4 a BN Av=|a, b,c) A,=|a, By 72 a, bs cy a Bs In terms of rows, we write R R A=|R) A =|R R R The multiplication process is analogous to the 2 * 2 case: RR RRL RR AA, =|RRI RR RR RR RR RRS Rows are multiplied similarly as before. For example, RARE = (a,52,56,) * (43, Bas¥2) = (G40, + By +C,72) Asin the 2 x2 case, we can have row-by-column and column-by-column multiplication, Let us evaluate the numerical product of two 3 « 3 determinants: 1 32 2 -1 3) A,=|4 2 -I| 4,=B 1 4 Is -1 3 o 5 2 Note that A, =-74and A, =15, so A,A, should be -1110. Now, let us evaluated 4,4, through row-by-row multiplication. We have, 13 2)(2 -1 3) RR; =(1 3 2) 1 4) RR, =(1 3 200 5 2) =14 =19 42-314) RRi=(4 2-110 5 2) 0 RR = (5-1. 32-13) RR=(5 -1.3314) —- RRV=(5 -1 30 5 2) =20 = 26 =1 42-12-13) RRy256 Determinants and Matrices So, 5 14 19 A,A,=| 3 10 8 |=-990+2198-2318 =-1110 20 26 1 3. Determinants and Linear Equations Consider a system of linear equations in two variables axtby +e =0 a,x+b,y-+e,=0 Graphically, these represent two lines which may be parallel or which may intersect at a unique point. If we solve this system, we obtain dd) a,b, | a, 2%, | If |2. &|#0, then a unique solution will exist for x and y. Graphically speaking, |% | # 0 means that a,b, ~a,b, # 0, or * # % , which implies that the two lines are non-parallel, so a unique solution must exist. aa ab | =0 thetwo lines are parallel, because # = %, Inthis case, if + # 2 = >, thetwolines are different and parallel, so no solution exists; this is'confirmed in (1) because the denominator- determinant is 0, but the numerator-determinants are not. The other case is when ==“ =}; thismeans that in 1), all the determinants will be 0. Physically, this means that the two lines are the same, so there will be infinitely many solutions. Let us now consider a system in three variables axtbhytez+d, a,x+bytoz+d, =0 axtbytez+d,=0 Physically, this corresponds to three planes in 3-D space. Ifthe three planes intersect in a point, a unique solution will exist for the system. But the three planes may not intersect in a point. For example, if two of the planes are parallel, there is no unique intersection point. In such cases, the system does not have a solution, Also, there will be cases where infinitely many solutions exist. For example, when the third plane passes through the line of intersection of the first two planes, the set of intersection points is the same line, which means there are infinitely many solutions.PART-A: ummary of Important Concepts 257 Let us now determine the solutions of this system. Taking a cue from the result of the two-variables system, we define a be ba a A=|a, 6, ¢,| and a,=|6, ¢ d,| a b ¢ bo 4, Now, on A,, if we apply the transformation C, + yC, +2C, + C,, we obtain D b,c byte,z+d, | nc byteztd, bo byteztdy But using the original equations, the third column C, reduces to bg -ax| lah 4, =|8, ¢ ~ax|= by bc, -ax a ho Using an exactly analogous process, we can arrive at expressions for y and z A ao d, *O. where A,=|a, cd a od a h& d A z=— where A,=|a, b, dy a Therefore, we arrive at what is known as the Cramer’s rule: Itis important to understand the justification we have presented for this rule (many students know the rule but not the reason why itis true). As is the case with a two variables system, if A #0, then obviously a unique solution will exist for (x, y, 2), given by the Cramer’srule. That solution will phy sically correspond to the point of intersection of the three planes represented by the three equations of the system, However, if A =0, then two cases may arise: in case all the other determinants A,,A,,A, are also zero, then the system will have infinite solutions. However, if even one of A,,\,,; is non-zero, then the system will have no solution, because Cramer’s rule will give rise to undefined quantities Let us summarize: Ao A=0 > Unique solution exists > If A,,A,,A, are all zero, infinite solutions exist > If atleast one of ,,A,,A, is non-zero, no solution exists A particular case of the 3-variables system is when all the constant terms are zero, Such a system is called a homogeneous system, since all the terms in each equation are of the same degree, i.., linear: axtby+ez=0 anx+b,y+o2=0 ax+by+oz=0 Note that x 0 isalways a solution of this system. But the system may also have infinite number of solutions, in which case there will exist solutions other than (0, 0, 0). The (0, 0, 0) solution is called258 Determinants and Matrices the trivial solution for this system, while other solutions (if they exist) are called non-trivial. For this system, A, =A, =A, =0, and so we have two possible cases: A#0 A=0 The system has the trivial solution (0, 0, 0) The system has non-trivial solutions, because and no other solution infinite solutions exist ference of matrices: If we have two matrices A and B of the same order m * n, their sum/ difference is given by adding/subtracting the corresponding elements A£B=[a,]+[0,) =[4, £5,] For example, 1 -2 201 30-1 30 4/+ )-1 3/=/2 7 5 3 42 9 -1 (b) Scalar multiplication: If we multiply a matrix 4 =[a, ] by A, we get the matrix 4 =[2a, ]. For example: fi 3 2)_[42 8 21-1] [8 4 -4 (©) Multiplication of matrices: To understand how matrices are multiplied, let us first consider arrow vector 7] ‘A row vector is a matrix G with just one row 4 and a column vector % |A column vector is a matrix with just one column “, which are both of order n, The product of R and C can be defined as G oo RCH=LF ry t)| F |= Ke, tne, ttre, & Therefore, RC is a scalar quantity. For example, 2 [1.3 2] }-1 4 Now, we will discuss matrix multiplication. To multiply two matrices A and B to find AB, the number of columns in A should equal the number of rows in B. Let A be of order m x n and B be of order nx p. The matrix AB will be of order m x p and will be obtained by multiplying each row vector of A successively with column vectors in B. Let us understand this using a concrete example a a, a, a B bb, by) B=la, By 6 4 % a; BsPART-A: Summary of Important Concepts 259 To obtain the element a,, of AB, we multiply R, of A with C, of B: Ga ad aa,+ eta, =] 2, 2 a | 3x3 3x2 32 3x3 3x2 3x2 Proceeding this way, we obtain all the elements of AB. In general, if A is of order m n, and B of order n x p, then to obtain the element a, in AB, we multiply R, in A with C, in B: NH | mxn nxp mxp és} (@) Identity matrix: For any matrix A of order m n, the identity matrix /is a square matrix of order n xn such that AI = A. If A is itself a square matrix of order n * n, then Al = IA = A In the identity matrix, all diagonal elements are 1 and all non-diagonal elements are 0. For example, the 3 3 identity matrix is 100 010 oo1 Ilustration 2: If 214 20 and B=|~-1 1|,find AB and BA 321 43 Working: Since A is of order 2 x 3, and B is of order 3 x 2, AB is defined and will be of order 2 2. BA is also defined, and will be of order 3 x 3. 20 w= [} 1 ‘] a -[? | 321) 4, 8 5 20 42 8 214 BA=)-1 1 =/1 1 3 3{l3 21 4260 Determinants and Matrices This calculation should have made it clear to you that AB is in general not equal to BA, that is, multiplication of matrices is a non-commutative operation. We note the following important facts about matrix multiplication: (@) For three matrices A, B, C, the equality AB = AC does not imply B = C. (b) Matrix multiplication is associative, that is A(BC) = (AB)C. Applications of Matrices to Linear Equations The application of matrices to linear equations is extremely important, and is therefore discussed in some detail here. Consider the system of equations Bxty422=3 2Qx-3y-z 3 xt2ytz= We can write this system in matrix form as 301 2Q)fx 3 2-3 -1 |] y |=]-3 12 1]lz 4 While studying determinants, we learnt how to solve linear systems using Cramer’s rule. We'll now see how to use matrices to do the same. Note that the matrix equation can be written as AX = B where 301 2 x 3 2-3 -1|, x=| y |, B=|-3 1 2 1 z 4 A simple linear equation ax =, where a,x, are reals, has the solution Can we do something similar for matrices? Can we define a matrix inverse so that X=A'B In case of square matrices, it turns out that we can! The inverse of a square matrix A should be another square matrix 4“! of the same order such that AA'=4'A Let us understand how to arrive at A” using the example of a 3 x 3 matrix: a by ey Note that the determinant of A, denoted as det(A) or | 4 |, by expansion along R, is, mal Dot spot a) lay c, lay, b, la, la, 4PART-A: Summary of Important Concepts 261 This can be written as a row-column product ‘We had also seen that the sum of co-factors of any row (or column) with the corresponding elements in a different row (or column) is zero. For example, lb cy la b,c, lb, cs 4, 1 (a; by; =O whereas [a,b ¢)]| + = [Alagain cs 45 ey b, la 5 by la; >, This suggests an interesting step forward. For each element d, ind, define its co-factoras C,.. Consider the matrix composed of the co-factors, but with co-factors of rows arranged as columns, For example, consider the matrix 4 4 a | with corresponding co-factors C,,Cyy.4Cy ‘Now, consider the matrix CCHEC! A=|C, & G Cmcmc! Note that ll 0 0 AA=| 0 ldl 0 o o ll Itis important that you are absolutely sure of this step. The matrix A is called the adjoint of A and is sometimes denoted as adj(d). Note that Ad is the same as 4, Thus, 100262 Determinants and Matrices ‘We have succeeded in evaluating the inverse of a matrix. Let us apply this to our original problem. 301 2 2-3 -1 121 Carefully observe each of the elements of A+ -1 A |, l4l=8 Since, we had AX = B, A'(AX)= A"B 3 1 => X=A'B 3]=| 2 4} [-l ‘We make the following observations’ * If |4| #0, 4” always exists which means that the system has a unique solution. © If |Al=0, two cases arise as in the case of determinants. Since, X = AB = 4B, (@) If AB # 0, no solution exists since ¥ becomes undefined. (b) If 4B = 0, the solution has an infinite number of solutions, * For a homogenous system, ie., B= 0, (a) If [Al #0, the system has only one solution, namely the trivial solution X= 0. (b) If |A|=0, then since 4B =0 too, the system has an infinite number of solutions. General Properties of Matrices 6.1 Standard Terminology (1) Zero matrix: (2) Row matrix: All elements of the matrix are zero. Also called a row vector, such a matrix has only row. Example: (1-12 4] Also called a column vector, such a matrix has only one column. Example: 1 3 -1 (3) Column matrix: (4) Diagonal matrix: All non-diagonal elements are zero, Example: 100 0-10 0 0 2PART-A: Summary of Important Concepts 263 (5) Upperilower triangular matrix: All elements below/above the diagonal are zero. 12-1 100 03 4 24 0 oo 1 -17 #1 Upper Lower Triangular ‘Triangular (©) Singular matrix: A matrix whose determinant is zero (7) Transpose: For amatrix A, the transpose of A, which we represent as, A? is obtained by interchanging the rows and columns of A. Example, 24 r[213 A=|1 1} > A= 417 37 Note that (4")" = 4. 6.2 Important Properties and Results (a) (4B)! = BTA’. In general, we have (4A, ... A,)” = AAT... Ay Al. Ap mmetric if A= A",and skew-symmetric if 4=-A". Every square das the sum of symmetric and a skew-symmetric matrix as follows: (b) A square matrix A i matrix M can be expres M= diem) + tiu—m") 2 2 Symmetric Skew Symmetric © “yt=a'y (@) A square matrix A is said to be orthogonal if 7A = 1 (i) The transpose of an orthogonal matrix is orthogonal. i) If is orthogonal, then |4|= +1. (iii) The product of two orthogonal matrices is orthogonal. (iv) The inverse of an orthogonal matrix is orthogonal, It will be a very useful exercise to prove these results about orthogonal matrices IMPORTANT IDEAS AND TIPS The Larger Picture. Why do we need matrices and determinants? Many of the readers might be surprised to know that matrices and determinants are an exceedingly important component of higher mathematic For almost all engineering disciplines, a thorough knowledge of this subject is required. The utility of matrices and determinants stems from the fact that they allows us to handle and operate on a large array of numbers simultaneously. This is particularly useful in solving practical problems whose mathematical models contain a larger number of unknown variables. Suppose that you are trying to solve a system of linear equations with 200 variables (you will encounter systems with a large number of variables in higher mathematics). Matrices and determinants give us a way to solve this system using a defined set of rules, We have solved three-variable systems, but the techniques can be applied to systems with any264 Determinants and Matrices number of varialbles. All we have to do is to program a computer to implement the various operations of matrices or determinants, and give the 200-variable system to this program as the input, Many computer scientists and mathematicians work on such problems. Mistakes in Evaluating Determinants, When evaluating determinants, students frequently calculate the co-factors incorrectly. The alternating + and — signs can lead to calculation mistakes. A particularly useful exercise to ensure that you never calculate determinants incorrectly is as follows: construct a few 4 x 4 determinants (select terms with small magnitudes) and evaluate them using a calculating utility (a calculator, or an online determinant evaluator). Every now and then, select one of these 4 x 4 determinants and evaluate it as fast as you can (by expanding along a randomly selected row or column), and then compare your answer with the correct answer. Doing this exercise frequently will minimize the calculation mistakes you might otherwise make. Mistakes in Matrix Multiplication. Mistakes are commonly made while multiplying two matrices, That is why we have explained matrix multiplication in detail in the preceeding pages, and you are urged to reread that discussion multiple times. If you have to calculate AB, the product of two matrices A and B, remember that (a) The number of columns in A and the number of rows in B must be the same (b) Rows in A multiply with columns in B to produce elements in AB. Each element in AB is formed by multiplying one row in A with one column in B. Left and Right Cancellation in Matrices. A very important concept to understand is that matrix multiplication isnot like normal (arithmetic) multiplication, For example, if Ihave the algebraic equation ab= ac, I can say that b = c. What I’ve done is the left cancellation of a, and doing that is possible in such equations. However, if [have the matrix equation AB = AC ,where A,B,C are matrices, then I cannot say that B= C.. That is, (left or right) cancellation does not hold in general for matrix multiplication. Scalar Multiplication, When amatrix is multiplied by a scalar, each element of the matrix gets multiplied individually by the scalar. However, when a determinant is multiplied by a scalar, only the elements of any one row (or column) may be multiplied by that scalar: a bc] [Aa ab ac lab cl laa Ab acl la 2b el Alp q rl=|ap aq ar| whereas alp q rl=|p 9 r| or |p ag rete x y z| lax ay az x y ale oy 2 x ay 2 Determinant Transformations. One numerical fact which you need to work out on your own is: can row/column transformations be clubbed together into a single step? For example, consider the determinant on the left, on which two row transformations are applied in the same step to obtain the determinant on the right la b el ROR +AR + UR, a+Aptux btAgtmy c+Ar+yz| BRP TT ROR HAR + OR ? i , ky a] BO RAAR TR Leap yrhbeig sthetsr Isit correct to do so? Secondly, can row and column tranformations be clubbed in the same step? Try working with some examples. Linear Equations. The most important application of matrices and determinants we have studied is related to finding the solutions of systems of linear equations, You must understand the geometrical significance of systemsof linear equations, and how matrices or determinants areused to solve them. Also, youmust appreciate how the two approaches are deeply connected. For this purpose, we summarize the results from the two approaches side by side. Compare the determinant approach with the matrix approach:PART-A: ummary of Important Concepts 265 (a) A two variables system: This system corresponds to two lines in a 2-D plane ax+by+e, axtb,y +e, Determinants Matrices (i) Problem Formulation a,b yx]_[-2 SPE or AX (ii) Case 1: The lines are non-parallel, Azo The inverse of A exists which means that there is a unique point of intersection, (iii) Case 2: The lines are different but A= The inverse of A does not parallel Atleast one of A,,A, exist and AB #0 is not zero (iv) Case 3: The lines are the same A=A,=A, The inverse of A does not exist and AB (b) A three variables system: This system corresponds to three planes in 3-D space axtbytez+d,=0 anx+b,y+e,2+d, =0 axtbyteztd,=0 Matrices ah ax] fod (@ Problem Formulation a,b, ¢||¥]=]-¢, a,b, e|lz] [-a, or AX-B (ii) Case 1: The planes are non-parallel, A20 The inverse of A exists. which means that there is a unique point of intersection. (iii) Case 2: At least two of the three A=0 The inverse of A does not planes are parallel, so there Atleast one of A,A,A, exist and AB #0. is no unique solution. is not zero, (iv) Case 3: The planes intersect in a A=A,=4,= The inverse of A does not line. This could happen if exist and 4B one plane passes through the intersecting line of the other two planes, or it could happen if two planes are the same and the third is non-parallel to them Alternatively, all the three planes could be coincident266 Determinants and Matrices Cramer's Rule. We have seen how the Cramer’s Rule applied to a system of linear equations helps us calculate the solution to that system by calculating the appropriate determinants, We would like to emphasize that you must be familiar with the justification behind this rule (for a system of 3 variables, we have presented the justification in the preceeding pages). In particular, you must understand that this tule is based on the transformation properties of determinants. Also, knowing the justification will help avoid calculation mistakes like taking the incorrect signs or incorrectly constructing the determinants. Inverse of a Matrix. You must remember that inverses exist only for square matrices, Secondly, you must be familiar with the justification behind the sequence of steps used to calculate matrix inverses. ‘We have presented this justification for 3 x 3 matrices in the preceding section. There is a very specific reason why the adjoint matrix is constructed, and you must understand that reason properly.Determinants and Matrices PART-B: Illustrative Examples Example 1 ha a’ The value of A=|1_ 6 b’Jis hee (A) (a-b)(b-ce-a)(a+b+e) (©) 2a-b\(b-ce-aatb +c) B) ~(a-b\b-cl(e-alatb+e) — (D) -2(a-b)(b-e)(c—aat+b+e) 1 R,>R,-R, Solution: We have A=|0 b-a Ba? * 0 - e-a RRR Expanding along C,, A=(b-a)(c -a’)-(b’ -a*)(c-a) -aye-a){(c* +a" +.ac)-(b +a? +ab)} =(a-bXb-c)(c-aya+b+e) The correct option is (A). . Example 2 ll+cos*@ sin? sin“? Let 0 € (0,4) satisfy | cos*@ 1+sin?@ — 4sin4@ |=0. The value of 244 is cos*@ —sin’?@— 1+ 4sin 49| A3 B85 ©7 Oil Solution: Using the transformation C,—>C,+C,+C, and factoring out the common factor (2+4sin 49) from C,, we obtain, 1 sin?@ sin 4a (2+4sin4@)|1 1+sin?@ — 4sin4@_ | =0. 1 sin?@ — 1+4sin40268 Determinants and Matrices Using R, > R,—R, and R, > R,—R,, we have 1 sin?@ 4sin40 2(1+2sin46)}0 1 0 0 0 1 Explanding along C,, we have 2(1+2sin40)=0 or = d-nst ek 6 6 The value of 244 is therefore 7. The correct option is (C). Example 3 Consider the homogenous system x-cy-bz=0 cx-y+az=0 bxtay-z2=0 If this system has a non-trivial solution, then the value of a* +b* +c” +2abe is (AO B1 Ol W)2 None of these Solution: If a homogenous system has non-trivial solutions, this implies that A = 0: 1 -c -b © -l a|=0 boa -l Using C,> C, +eC, and C,> C, +6C,, we have 1 0 0 Aslc -l+#c? atbe |= b atbe -1+b => (-1+c?)(-1+b7)-(a+ be)? =0 Upon simplifying, we have a? +b? +c? +2abe =1 Thus, the correct option is (B). Example 4 Let A, B, C be the angles of a AABC. What is the value of the following determinant? eM et ow Anle® oe et ee pt gn (AO B)2_C)4_ — D)8_— None of thesePart-B: Illustrative Examples 269 Solution: We'll make use of the fact that 4 +B+C =n, Taking out factors of e4,e",e~ from C,,C;,C respectively, we have: 1 1 1 Aa e280] plese gi25c) QKB-A eriPaem gilc-a)gs20) Now, we note that 24+C =n -(B-A),2B+C =n-(4-B),2C+B=n-(A-C),and 244+B=n+(A-C). Therefore, 1 1 1 trot aaerno|_g-n yn gan |_|_y 1 y]og elena ge _gte-w | |p yy The correct option is (C). Example 5 Let f,g,h be differentiable functions of x and A and be such that f g h A=] GA Gs) hy’ CeA" Gg) On" We also define two more determinants as follows: fogeh f gh a-lf et i | and A-| fog yg it yn gt i Which of the following expressions for 4’ (the derivative of 4") is correct in terms of A, and A,? (A) A’ =2xA, 43x70, (C) Aa PA HAA, ®B) A'=x°A,+2xA, () A’ =3x7A, +A, Solution: We have f 8 h A= feof gtxg' h+xht Uftedaf'+2f xg" +4xg'+2g Ch" +4xh' 42h Applying R, > R, -R,, R, > R,-4R, +22, we have foghk A=x|f' gi h'l=xA, yet ht => A'=3x7A, +07)270 _ Determinants and Matrices where fs wy|f g hl |f gh ately et later et alee et at yg [eet wl pt gt at =0+0+A, => A'=3x7A, +2°A, ‘The correct option is (D).Part-B: SUBJECTIVE TYPE EXAMPLES istrative Examples 271 Example 6 lloga p 1 Let a,b,c be the pth,qth,rth terms respectively of a GP. Evaluate A=llogb q 1) loge r | Solution: If we let the first term of the GP be fand the common ratio be ¢, then we have a=fU" = loga=log f+(p-llogt. Similar relations can be written for log 6 and log ¢ logf+(p-Ilogt p 1| |(p-Nloge p 1 A=|logf+(q-Dlogt gq 1]=| (q-Nlogt gq 1| C,>C,-(logr)c, logf+(r-Ilogt r 1{ |(r-Iloge r 1 p-l pl Ppl =logt}q-1 q 1|=logelg g 1| C+C+G=0 rtrd rod a Example 7 Evaluate ax-by-cz aytbx — extaz A=| aytbx by-cz-ax bz tey extaz — bztey ez ax-by Solution: Let us try to reduce C, to a simpler form. By the transformation C, > aC, +bC, +eC,, we generate the factor (a’ +4” +c’) in all the terms of C,: x(a’ +b +c") aytbx ex+az (Note that there is a factor of 1) ya +B? +c7) by-cz-ax bz toy 9| (at +b +c?) betcy — cz-ax-by x aytbx — cxtaz (a? +b? +c*)| y by-cz-ax — bzt+ey z betcy — cz—ax-by Using a similar artifice, we can simplify R, now. By R, > xR, + yR, +2R,, we have xeytz? boty +z?) c(t ty? +z") (+B +c)} py by-cz-ax betey z betey cz-ax-by 1 b c (@ +b +7 ty +z°)| y by-cz-ax bztey z brtqy cz—ax—by272 Determinants and Matrices Now, we simplify C, further by the row transformations R, > R, — yR, and R, > R,~ 2R;: 1 6 e A Leis ect tye) 0 -cz-ax bz ax 0 cy -ar-by Finally, we expand along C, and simplify A=(@ +B +0°\x' ty? +2" art by+cz). a Example 8 Evaluate (a= bY (@q- by (a,-b,) A=|(a,-B) (a,-bY* (a,b)? (a, -bY (a-by (ab) Solution: The trick is in recognizing the fact that A can be expressed as a product of two 3 = 3 determinants A, and A,. To arrive at this product, we note how elements of row 1 can be expressed as products of rows, (a,-b) =a; —2a,b, +8) = (a? 2a, I)x(1 -b, 8) (a,-b,)’ = a} -2a,b, +b; =(aj 2a, 1)x (1 -b, 63) (a, -b, =a} ~2a,b, +53 = (a> 2a, 1)x(1 -b, BF) This means that we now know how to fill R, of A, and the three rows of A, a@ 2a 1 1 -b bP A=|— 2 >|x]1 -b, BF | 2 3] [1 -b BF a 2a, 1 A=|a; 2a, 1|x @ 2a 1 These two determinants are separately very easy to evaluate; the produet comes out to A= 2a, ~ a, )(a, ~ a, )(4, ~ a, )(b, ~ B, Mb, = by Wd; =). a Example 9 (a) Prove that the system of equations Bx-y+4z=3 x+2y-3: 6x+5y+Az 3 has at least one solution for any real 4. (b) Find the solution(s) for a =-S.strative Examples 273 Solution: We have 3-1 4 A=|1 2 -3/=7(4+5) 65 a If A#-5,A0, soa unique solution exists. For A =—5, we have -1 4 33 A,=|2 -3 2|=0 5 -5 3 Similarly, A, = A, =0. Thatis, for 4 =-5,4 solutions exist. Therefore, in all cases, the system has at least one solution. A, =A, =0, so in this case infinitely many ‘Now, how do we express the set of infinite solutions in the case of 2 = -5? What we have to realize is that if we express two of the variables in terms of the third, and vary the third as a parameter, we get all the possible solutions, For = 5, observe that y and z can be written in terms of x as 3x x —— (Verify! 3 (Verify!) mS Asx is varied, y and z vary in a manner such that (x, y, 2) is a valid solution to the given system. Therefore, the set of all possible solutions can be expressed as =x 4- ( x 4 7) rer 5 Example 10 Find the value of (a -BY (q- bY (a,b (a,b) (a, -b) (a, -bY (a, -bsY° (a, -b,)° (4-6)? (ab, (a-BY (a,b, (ay-b)Y (ay—by (a~bsY° (ay)? Solution: The given determinant can be expressed as product of two simpler determinants a -2a, 10 16 BO a-[@ 2m 10), | & Bo}, a -2a, 1 0 1b BO a, -2a, 1 0 1b OO Example 11 Prove that ca-b? ab-c 2 2 P ab-c’ be-a@|=|q p ql, be-a’ ca-b’| |q where p=a? +b? +c? andq = ab+be+ca,274 Determinants and Matrices Solution: If you consider the determinant on the LHS, say ,, carefully, you'll see it is the determinant of the co-factors of (the determinant on the RHS) . Example 12 Let A= f 2 A ] Evaluate (AB) C and A (BC) and verify that they are the same. Solution: The orders of A, B, C are 2 x 3,3 x 2 and 2 x 4 14 123 13-7 aa-[} 1 a at [? i . 23 13-7) [2 1 4 3]_[33 13 31 32 (ABYC = 6 Wfl[-l 031 1 6 57 29 14 -2 1 16 7 2143 Bc=|3 -1 =|7 3 98 o31 23 72-13 v2 aq {2 1 8 7) 33 a3 a1 32 A(BC) = 73.98 3.10 1 6 57 29 7 2 -1 3. As expected, (AB) C and A (BC) are the same. . Example 13 Consider the system of equations xtyt+z=5 xt2y432=9 xe3ytazay Find and j1 for which this system has (a)aunique solution —_(b) no solution _() infinite solutions Solution: We have hon u = 1Part-B: istrative Examples 275 (a) Fora unique solution, A¥0 => A#5, (b) Forno solution, A =0 and AB #0.Now, A=0 meansthat 2 =5.For 4 = 5, the coefficients matrix A becomes rid 12061 A=|1 23] > A=|2 4 2 135 1201 1-2 17/5 H-13 => AB=|-2 4 -2)| 9 | =| -2u+26 12 ile H-13 Since 4B #0, #13. Therefore, the system will have no solution if 4 = 5, #13. © For infinitely many solutions, A=0, AB=0 => A=5, w=13. . Example 14 For two matrices A and B whose product is defined, prove that (4B)" = BT 4”. Solution: This general result is extremely important, and so you must understand its proof carefully. Let A be of order m * n and B of order n x p. Then AB is of order m x p. Now, A’ is of order nx mand B’ is of order pxn. So B’ A’ is of order p x m which is the same as the order of (AB)’, Let us understand this using a concrete example: ed [ts cd xyz => (4B) PT egtby artbe cptde cqtdy ortdz ap+bx cp+de => (AB) =| ag+by cq+dy ar+bz ert+dz Also, px ‘ap+bx ep+dx Ba =|q y [; ‘J- aq+by cq+dy rz arsbz ortde Focus on a particular term, say ap + br, and note how it is generated at the position in both (AB)! and B’ A’. Now let us generalize this. Consider the term in (4B)’ at the position (i, /), say In AB, this same term is at the position (j,i). So, ris generated from R, in A and C, in B Let us now consider the term at the position (i, /) in B”4". This will be generated from R, in B” and C, in A", or C, in Band R, in A, which will generate ¢ again. Thus, (4B)' = BTA" a276 _ Determinants and Matrices Example 15 Asquare matrix A is said to be symmetric if A = A”, and skew-symmetric if A =—A". Show that every square matrix can be expressed as the sum of a symmetric and a skew-symmetric matrix. Solution: If A be the given matrix, then A can be written as 1 A gata’) + 54-4) Symmetric Skew-symmetrie (®) A+A" is symmetric because (A+ A")! = AT +(AT)’ = ATH A Gi) A-AT skew symmetric because (4- AT)! = AT - A=-(A- A’). For example, let 1-1 0 A=|3 2 6 24-3 ‘Then, 111 0-2-1 54ea%)=] 1 2 5) and Y=—(4-47)=|2 0 1 53 1-1 0 Note that ¥+Y= 4 7 Example 16 Let A be a skew-symmetric matrix. Let B be a matrix of the same order such that B(/ - A) = 1+ 4. Show that B is orthogonal. Solution: We have to show that B’B = BB’ = J. Note that since A is skew-symmetric, A" = —A. We are given that BUI-A)=1+4 ay ‘Taking the transpose on both sides, we have (I-A)! BT =1+A" => (I+ AB" =1-A Q) From (1) « (2), we have (1+ A)B"BU— A) = (I~ A+ A) = (1+ A) A) (How?) => B'B=I By (1) * (2) again in another order, we have BU~ AML + AB =P -# => BI? -A)B'B=(2-A)B GB)lustrative Examples 277 Since BTB=1, we have BUP -#)=(P -4#)B @ Using (4) in (3), we have (P -A)BBT =P -# => BB" Thus, BB" = BB =I. 7 Example 17 Let A be the set of all 3 x 3 symmetric matrices all of whose entries are either 0 or |. Five of the these entries are | and four of them are 0. (a) Find the number of matrices in A. (b) Find the number of matrices in A for which the system of linear equations x] fi Al y |=] 0 z} lo has a unique solution (© Find the number of matrices in A for which this system of linear equations is inconsistent. Solution: (a) Since all the matrices are symmetric, this means that on the diagonal, there can be either one ‘I’ or three I's. There cannot be two ‘I's. If all entries on the diagonal are 1: In this case, 3 matrices are possible, since the other two ‘I’s can be placed in 3 possible ways. If one entry on the diagonal is 1: In this case, 3 x 3 = 9 matrices are possible, since the ‘I’ on the diagonal can be placed in 3 ways, and the other two pairs of ‘1’s can be placed in three ways. => The number of matrices in A is 12. (b) For a unique solution, | 4 | should be non-zero, We count such cases explicitly (@ All diagonal elements are 1 xy 12 | One of x,y, zis 1, the other two are 0. z1 0, then |4|=0. Verify this explicitly. Similarly, if y = 2 =0 or if x = then also [4|= 0. So in this case, there are no matrices for which a unique solution exists One diagonal element is 1, the other two are 0 Ixy Oxy Oxy x 0 z}or |x i zlor|x oz yz 0 yz 0 yrl278 Determinants and Matrices In each case, one of x, y, zis 0, the other two are 1. In the first case, if x= 0, y=2=1, then |A|#0. Ify=0,x=2= 1, then |4|#0. If2=0,x=y= 1, then |4|=0. Thus, there are 2 matrices from the first case for which | A|+ 0. Similarly, each of the other two cases give rise to two matrices with non-zero determinants => There are 6 matrices 4 for which |4|#0_ (for which the system has a unique solution). (c) There are 6 cases where |A| = 0. Out of these, in 2 cases, namely, tii 100 10 0} and Jo 1 1], 100 oil the number of solutions is infinite. The other 4 cases are inconsistent systems aDeterminants and Matrices PART-C: Advanced Problems PL. P2. P3. PA, OBJECTIVE TYPE EXAMPLES Let a, b, ¢ be real numbers with a? +6? +c? =1. What does the following equation represent? ax-by-c bx+ay cxta bxtay -axtby-c cy+b | =0 exta cytb — ~ax-byte (A) Apoint — (B) Astraight line (C) Acircle (D) Aparabola._() None of these sinx — sin(x+#) sin(x+2A) If A=|sin(x+2h) — sinx —_sin(x+A)], the value of lim, (3) is sin(x+h) sin(x+2h) sinx (A) 6sin xcos* x (B) 6cosxsin? x (C) 9sinx cos? x (D) 9cosxsin? x Suppose that /(x) is a function satisfying the following conditions: $.and (¢)Forall x, (a) f(0)=2, fQ)=1 (b) fhas a minimum value at * 2ax 2ax-1 — 2ax+b+1 S@O= b b+l -1 QAax+b) 2ax+2b+1 2ax+b where a, b are some constants The value of (3) is 1 1 H+ @t ©1 M2 G oh a oP © @) A2x 2 real matrix is given such that 4A” = J. What is the maximum value of the sum of the absolute values of entries of matrix A? A BV C22 — WyNone of these v2 Let A, B be skew symmetric matrices. Which of the following are skew- (A)A+B @)AB_(C)AB+ BAD) AB-BA_—@) A" symmetric?280 Determinants and Matrices Po. P7. P10. Pu. Let A=| < «| such that [4|=0. Ifa, b,c are distinct, then the sum of the coordinates of the fixed point through which the line ax-+y +e =0 passes is A2 B83 ©4 Os Let us define the power of a matrix A as the minimum m € Z* such that A" = 1. For two matrices A and B, if A’ =1 and ABA™ = B?, the power of matrix B is between (A)20and24 — @) 28 and 32. (C)36and 40D) 44 and 48 Let A be an x n matrix with |A|=4. B is the adjoint of the matrix 24 such that |B|=1024, What is the value of n? (3 B)4_— ©)S_—@D) More than 5 Let 4=|sxf, s38] and B=(, |], and define matrix C equal to (AB AT)" where » €Z*, The minimum value of the product of entries of C is 1 ©2 —— @Noneof these 5 5 v2 1 als (D) — Q~ OF Let A and B be two non-singular matrices such that (4B)' = A‘ B* for three consecutive positive integer values of k. Match the matrices in the first column to the corresponding matrices in the second column. (A) ABA" Pa (B) BaB* QB (C) AB’A' RA (D) BAB" (S)B If Mis a3 «3 matrix, where M’M = and det (If) = 1, then the value of det (Md J) is (A)1 (B)2_—(C)4_— (DY None of thesePart-C: Advanced Problems 281 PI2. P13. Pu. P15. P16. Pi. PIs. SUBJECTIVE TYPE EXAMPLES 8s S55; SHS If S, =a‘ +b‘ +c’, evaluate Consider the following facts: a cows graze b fields in ¢ days 4a, cows graze b, fields in ¢, days 4a, cows graze b, fields in c, days Assume that all the fields provide the same amount of grass, that the daily growth of the fields remains constant, and all the cows eat the same amount each day. Find the value of 6b ch ca Ce lb, eb, ¢0,) Let (J,, m,, 7,),i=1, 2,3 be the direction cosines of 3 mutually perpendicular lines, Find the values of @ LLM, L Lm. mn, Dh, For any value of @, find the value of the following determinant: sind cos sin20 sinf+2%) — cos(@+2) —sin{ 20 + 2% 3 3 3 sin(-24) cos{-24) sin( 20-42) 3 3 3 If a,d,c,x,y,z € R, then prove that (a-xP (b-xP (e-x)'| tax)? (+ 6x)? (+ ex)? (a-») (b-y? (e-y)'] = [d+ am) tb +e)" (a-zy (b=zy (e=2z) (l+az) (l+bz)? (1+ez)* If f(x) isa polynomial of degree <3, prove that 1g a sal lpg @ te lg ele LO) x-b 2 | G=ayx=bye=0) fle lee le x-c For what values of p and q does the system of equations 2x+ py+6z=8 x+2y+qz=5 xtyt3ze4282 Determinants and Matrices P19. P20. P21. P22. P23. P24. P25. P26. P27. (a) have no solution? (b) have a unique solution? (©) have infinitely many solutions? For a matrix A, define the eigen-vector as a non-zero column matrix. and eigen-value a scalar 4 such that AX = AX. (a) Prove that the eigen-values of a symmetric matrix are real. (b) Prove that the eigen-values of a skew-symmetric matrix are 0 or purely imaginary. (©) Find the absolute value(s) of the eigen-values of an orthogonal matrix. A matrix Mis said to be orthogonal if MM" = M"M = 1. Let Sbe a skew-symmetric matrix and / the identity matrix. If [— Sis invertible, can we say that the matrix. A = (/ +S) (IS) is orthogonal? Let U,,U;,U, be column matrices satisfying AU, = AU, AU; =A,U;, AU, = BU, where is another matrix, Find A Let A, B be square matrices of order n, Can AB ~ BA be equal to the identity matrix? Let S'be the set of invertible matrices satisfying the following properties. (i) A-I=A forevery AeS (ii) For every Ae S, there exists Be S such that 4-B=1 Show that (a) B-A=I (b) 1-A=A Let A be the matrix [!, |} (a) Show that 4? -24+2/ =0. (b) Hence evaluate the value of B given by A’ ~ A+ 2A — A+ 4A°-24 44d + AAP + AGL. Let A and B be distinct 3 x 3 matrices such that 4° = B° and 4°B = B74. Can we say that the matrix A’ +B? is invertible? Let 4=[a,,] be a square matrix. Define by the conjugate of A the matrix A=[d,,]. A square matrix His said to be Hermitian if M =(1)'. Let A be a Hermitian matrix written in the form 4 = P+iQ where P, Q are real matrices. Which of the following statements is true? (a) P is symmetric (b) Qis skew-symmetric. Consider a 2 2 matrix A =|? q Consider a polynomial P(A) generated by the equationPart-C: Advanced Problems 283 P28. (a) Show that P(A) =0 (b) Define e* through the following equality ef arta 2 4 fen Show that e* can be expressed as a4+b/, where a, b are constants, (©) Determine e* for Oo x (ii) _x 2x Let C be the matrix defined as 4 a, a 4, a, a a, 4 CH] on Oy G2 a, a a, 4, Let f(x) =a, +a,x+a,x? ++-+a,x"". Let a, value of Cin terms of f(a,),i=1, 2,...1. ton be the n* (complex) roots of unity, Find theDeterminants and Matrices PART-D: Solutions to Advanced Problems S1._ In this problem, we will only be providing the sequence of steps, and you can work out the details yourself. Step 1: Ri > aR, R, > bR,, Ry > eR, Step 2: R, > R, +R, +R, and use a? +b? +c7 Step 3: R, 9 4,R, 9% Step 4: R, > R, -bR,,R, > R,-cR, Step 5: Cy >C,-x€, + Step 6: Expand along C, to obtain (x” + y*) (x+ by +c) =0. This is (apart from the point (0, 0)), the line x+by-+c =0. The correct options are (A) and (B) S2. We apply the following transformations to A in sequence: C, > C,-C,, C, > C,-C, C, > & and C, > | Now using lim,,,, on the resulting determinant, we have sinx osx cosx Ly a lim =| sinx -2cosx cosx I a? sinx cosx -2cosx 3sinx 0 0 R>RAR +R, =|3sinx 0 -3cosx RR, +2R, sinx cosx —2cosx =9sin xcos? x The correct option is (C). S3. Applying the transformation R, > R,—R, -2R, on the given determinant reduces it to 2ax 2ax-1 2ax+b+l S@)=| 6 b+1 -1 |=2ar+5 o 0 1 > fa)eat+brre From the first two conditions, a, b, c, are easily determined, and we have f(x) Thus, (3) = +. The correct option is (B).Part-D: Solutions to Advanced Problems 285 S4. Ss. 6. s7. Let a-[? ‘] =[f al: a cd c djlb d = a+b =e +d? =1and ac+bd =0. Solving these tells us that A will be of a form similar to ("2 2°] (verify). The sum of absolute values, 2(|sin@ | +| cos@ |), has a maximum value of 2/2. Proving this is straightforward, and left to the reader as an exercise. The correct option is (C) (A) (A+ BY" = A’ +B" =-(A+B). This matrix is skew-symmetric. ®) (4B) = B’ A’ =(-B) (-A) = BA This matrix is symmetric. © (AB + BA)! = (AB)! +(BA)" = BT AT + ATBT = (-B)(-A) + (-A)(-B) = BA+ AB. This matrix is symmetric @) (AB BA)" =(AB)’ -(BA)’ = BTA! — AB = BA- AB. This matrix is skew-symmetric My yt ay =a =f 4 now -A" if nodd Thus, A" is skew-symmetric only if 7 is odd, and not in general We see that the correct options are (A) and (D). A upon simplification is (a+6+c) (a—b) (b-c) (c—a). The only way this can equal 0s if atb+c=0>a(l+b()+e=0 The fixed point is therefore (1, 1). The required sum is 1+ 1 = 2. This means that the correct option is (A). Wehave B* =(ABA™') (ABA) = ABA" = A(ABA")A" = A?BA* => BY=ABA* => BY =A'BA* => BY =ABAS=B => B'=I Thus, the power is 31. The correct option is (B).286 Determinants and Matrices $8 B=adj(2A) = |B\=|24\" (why?) => 1024=4"'2""" (again, why?) = agrer => n=3 The correct option is (A). $9, Wehave 4d’ =I, ie, A’ =A". Thus, C= (ABA")" =(ABA")(ABA")---(ABA") = AB" AT _[ cos@ sind] [1 »][cosé -sino “| -sin@ cos@ | [0 1 || sind cose _| l+nsin@cos@ nos’ =nsin? nsin@ cos => P(product of entries) = -n’ sin? @ cos’ @(1—n’ sin? @ cos” @) v2 > | ot sin20 4 The correct option is (B). $10, Thisisa very interesting problem on the commutative/non-commutative behaviour of matrices, Specifically, we always need to keep in mind that matrix multiplication is in general non-commutative. Let (AB)* = A'B* hold true for k =n, n+1,n+2, where n eZ. Thus, (a) (AB)" = A"B" = (AB)(4B)" = A""'B""' (using(a)) (b) (4B) = 4a™'B | => (AB)A"B" = A""'B"' —(left-cancellation of 4; (©) (ABy"? = A™?B"? => BA" = AB right cancellation of B") We therefore have BA" =A'B () Similarly, using (b) and (c), we will have Bat aa BR Q) Thus, BA™ = A™"'B => (BA")A=A™B => A°BA=A""'B (using(1)) => BA=AB (left-cancellation of 4”)Part-D: Solutions to Advanced Problems 287 Su. We have managed to show that A and B ate commutative. Now we proceed to the actual question(s) (A) ABA" =(AB)A" =(BA)A™ = B(AA") =B (B) BAB" =(BA)B" =(AB)B™ = A(BB")= A (C) ABA" =(AB)BA" =(BA)BA" = B(AB)A* B(BA)A* = BB(AA* (D) B4’B" =(BA)AB™ =(AB)AB" = A(BA)B" = A(AB)B™ = AA(BB") = A? ‘The correct matching is (A) to (Q), (B) to (R), (C) to (S) and (D) to (P). We have (M-D! =MT-1 Multiplying by MZ, we obtain (M-D!’M=M™M-M=I-M =-—(M-I) Taking the determinant on both sides, we have |(M-D! M|=|-(M-D| => |\(M-D"||M|=-|M-1| => |M-1|=-|M-1| = |M-1|=0 None of the first three options is therefore correct, so (D) would be the correct response.288 Determinants and Matrices S12. S13. sid. SUBJECTIVE TYPE EXAMPLES Split the determinant into a product SS SJ flor 111 => A=|S, S, S,J=|a 6 c]x]a 6 c| (How? 5&5, S| la Bel la Be = {(a-by(b-ee-a)y? Let the initial quantity of grass per field be P, the daily growth per field be Q, and the grazing capacity of each cow be R per day. We thus have b(P+0Q)=c-aR b be ca B(P+eQ)=c-aR = |b be em b(P+6,0)=¢,a,R by bye ray We let Lom om As|l, m,n; ‘my my Since [? + for i=1,2,3, while 1, +mm,+nn,=0, where i,je(l2,3},i4 j, ‘we conclude that njfh bh &] fl 00 Ad’ n||m,m, m,|=]0 1 0 a\lm om ny} [0 01 Sinee 44” =I, we have (AAJ =I =1 = ATA=I Thus, Ab |L hom m, m, m,|\1, m, mm, My Lom ye Lim, 0 => |Xim, Ym 0 Yin, Sma, 1 m2 =n: _, [DPD -d Sim SinPart-D: Solutions to Advanced Problems 289 S15. We simply apply R, > R, +R, +R, to obtain all elements in the top-row as 0, and thus A = 0. S16, This is an interesting problem on determinant row/column operations. The trick is to write the left determinant A, as a product: lx x 1 2a @ A=]l y y? |x {1 2 B? lz 2 1 2¢ ¢ and now multiplying ‘row-by-row’, from which the right determinant A, is immediately obtained. We leave it up to the reader to figure out the details. S17. We construct a new function /(x) as follows: x-a a(x-a) f(a) laa Mx)=|x-b (xb) f(b)|- f(b M x-c ¢(x-c) f(c) lec Why did we construct this particular function? Well, the final result we need to prove has guided us, and the reader is urged to make the connection, especially after going through the entire solution. Now, we note that A(x) is again a polynomial of degree <3. The essence of the solution is in this very fact. To proceed, we evaluate (a): 0 0 S(a) lad ha) =|a-b Bla-b) f(b)|- f(a]. b a-c cla-c) f(c) lee = {ela bya —c) —b(a—bla—e)} f(a)— f(ay(a—bb—cNle—a) =(a—b)(b—c)e—a) f(a) —(a—b)(b— ee a) f(a) =0 Similarly, (6) = h(c) = 0. What can we infer from this, given that (x) is at the most a second degree polynomial? Well, since a second degree polynomial cannot have three zeroes, we must have the coefficients in h(x) to be all zero, so that h)=0 V xeR Using (1) and slightly rearranging, we obtain the desired expression we were required to prove. S18. We evaluate the appropriate determinants: 2 1 3-q\(2= p) 1 Hes 6 q 3 If p+ 2and q #3, then the system definitely has a unique solution. If p =2 or q =3 orboth, the system will have no solution or infinitely many solutions, depending on the values of the other determinants290 Determinants and Matrices 8 p 6 3 2 q|=(5-49)2-p) 413 286 A,=[1 5 q]=0 (regardless of the values of p and q) 143 2 pes A,=|1 2 5]=p-2 114 ‘The answers are: (a) p#2,q=3 (b) pP#2,q43 (©) p=2,q canhave any value. $19. (a) Since A is symmetric, we have A= A", while A= A because A is real. Now, AX =AX = XTA=AXT => XTA=AXT => RAX)=TX'X > AK X)=ARX) XTX represents the inner product of X and X7, and must be non-zero for non-zero X (why). Thus, A= = Theeigen-values are real (b) In this case, A’ =—A: AX =AX => -XTA=AXT = -X'A=TXT => -X(AX)=AXX => -A(XTX)=A(¥X) This means that A+4=0, i., A=OorkiforkeR => Theeigen-values are 0 or purely imaginary. (c) IfA is orthogonal, then 4’ A = AA” =I: AX =AX => XTAT=AXT => XTAT(AX)=AXT(AX) => X(ATA)X=A(XTX) => XTX =V(XTX) This gives 27 =1 => The eigen-values have absolute value 1 S20. We have A’ A=((1+S)\U-S)') +S) -Sy" Since the transpose of a product is the product of the transposes in the opposite order, we havePart-D: Solutions to Advanced Problems 291 A’ A=((1-S)')' (+S) (+S -sy" =((I-S)' (+87 1+ S\I-sy" =(1+S)"(U- Syl + Sy sy" a We observe that J-S and J+ must commute (why?). Thus, we switch the orders of the second and third factors in (1); A’ A=(1+S)'(U+S\1-S\I-Sy! (2) Finally, it is given that /—S is invertible, and as a consequence, its transpose ({—S)" = J +S must also be invertible: (U+sPU+s } ee (-S\u-sy' From (2), we have A A=I Similarly, 44” =] can be proved, ie., A is orthogonal S21. Let Ube the matrix with columns as U,,U,,U,. Then, AU = AU, U, Uy)=[4 U, 4 Uz A U5) 4 0 0 =U,U,U,]] 0 4, 0 004% Denote the second matrix on the right hand side by D. Then, we have AU =UD => A=UDU* S22, The diagonal elements of AB and BA are the same, so AB ~ BA will have its diagonal elements 0. Hence, it cannot be equal to the identity matrix. $23. Let Ae S. There exists a Be S such that 4-B=J. Further, for B, there exists a Ce S such that B-C=1.Now, B-A=B-A-I (Property (i) =(B-A)-(B-C)=B-(A-B)-C=B-I-C=B-C=1292 _ Determinants and Matrices Further, AB=I=> A-BA=I-A => A(B ASA > AT=I-A=A S24. Part (a) is trivial. To obtain B, consider S(x)= ax" x9 4208 x7 44x’ 2x8 det to a ttl and divide fx) by x° -2x+2 to obtain a quotient and a remainder: F(x) = (8 = 2x42) (8 + x7 2x8 4° 4 Ox tt Dt x=1 Now, substitute x= 4 to obtain 01 f(A)=B=A-1 [S ‘| $25. No. We have (4? + B*)(A~B)= 4’ ~ 4°B+ B°A~B’ =0. Since A and B are distinct, this implies that 4° + B? is non-invertible, for if it were invertible, then assuming the inverse is C, we would have C-(4° +B’) (A-B)=C-0=0 => 1-(4-B)=0 = A=B, acontradiction $26. Since A is Hermitian, A = (A)’- => P+iQ=(P+iQ)' =(P-iQ)' =P" -id” => P=P", Q=-Q" Thus, both the statements are true. 827. (a) We have P(A) =0 => M-(a+d)At+ad-be=0 => P(A)=A’-(a+d)A+(ad be) _| @+be ab+bd ]_[ a’ tad ab+bd 4[ dee 0 acted be+d? | | acted ad+d? 0 ad-be =0Part-D: Solutions to Advanced Problems 293 (b) The crucial step here is to note that since e* isa polynomial, and A satisfies a polynomial equation of degree 2, we can write e* = P(A)Q(A)+ R(A), where the degree of R(4) is less than 2, that is, R(A) is of the form ad + bf. Since P(A) = 0, the given assertion is true. (©) Gi) P(A)=A? =x°. Thus, e' =(A?-x°)Q(A)+ a+b, ie., e' =ax+b and e =-ar+b = a=£—= and p=£** 2x Hence, e 0 ef =aA+bI = (i) P(A)=2? +2ax+¥° =(A+x)*, Thus, 8 =(At+ x Q(A)+ ah +b a Letting 2 =—y, we have e* = ~ax +b. Differentiating (1), and letting 4 =—x again, we have => baxe*+e*=(x+Ne™ => ef =ad+bl = als * ]294 Determinants and Matrices Now, we evaluate the produet C 1" (a) FG) F(,) cy-| @/@) af las) a f(a) a" f(@,) . ari f(a,) = SG) F (Gq) f (4,)V = C= Sf) S(@,)..f(G,) Note how I”is generated again in the product, soit cancels out. This artifice is what makes the problem very interesting! $29, Let fay|>la,| + lash, [bal > 18,1 + [Bl and fo,|> le, + le, and let 4 4 4 A=|b, bb Ga & Can A be equal to 0?