mel The Solution of Linear Simultaneous Equations Circuit analysis frequently involves the solution of linear simultaneous equations. Our purpose here is to review the use of determinants to solve such a set of equations. The theory of determinants (with applications) can be found in most intermediate-level algebra texts. (A particularly good reference for engincering students is Chapter 1 of E.A. Guillemin's The Mathematics of Circuit Analysis (New York: Wiley, 1949], In our review here, we will limit our discussion to the mechanics of solving simultancous equations with determinants. A.1 Preliminary Steps ‘The first step in solving a set of simultaneous equations by determinants is to write the equations in a rectangular (square) format. In other words, we arrange the equations in a vertical stack such that each variable occupies the same horizontal position in every equation, For example, in Eqs. A.1, the variables i, is, and i; occupy the first, second, and third position, respectively, on the left-hand side of each equation: Diy = Siz 12i, = —33, =i + 64; = 25, = 3, (aay Bi — Ai, + 22iy = 50. Alternatively, one can describe this set of equations by saying that i occupies the first column in the array, i; the second column, and i; the third column. If one or more variables are missing from a given equation, they can be inserted by simply making their coefficient zero. Thus Eqs. A.2 can be “squared up” as shown by Eqs. A.3: 20; — 0, 4u; + 303 = 16, (az) Toy + Quy 2vy — vy + Ovy = 4, On, + dvy + 305 = 16, (aa) Tey + Ovy + 205 = 5. m5126 ‘The Solution of Linear Simultaneous Equations A.2 Cramer's Method ‘The value of each unknown variable in the set of equations is expressed as the ratio of two determinants. If we let NV, with an appropriate subscript, represent the numerator determinant and A represent the denominator determinant, then the kth unknown xj is Ne ast aa ‘The denominator determinant A is the same for every unknown variable and is called the characteristic determinant of the set of equations. The numerator determinant N, varies with each unknown, Equation A is referred to as Cramer's method for solving simultaneous equations, A.3_ The Characteristic Determinant ‘Once we have organized the set of simultancous equations into an ordered array, as illustrated by Eqs. A.1 and A.3, it is a simple matter to form the characteristic determinant. This determinant is the square array made up from the coefficients of the unknown variables. For example, the characteristic determinants of Eqs.A.1 and A3 are a -9 -12 A=|-3 6 2 (as) “8-4 22 and 2-1 a=|o 4 3], (a6) 7 0 2 respectively. A.4 The Numerator Determinant ‘The numerator determinant N, is formed from the characteristic determi- nant by replacing the Ath column in the characteristic determinant with, the column of values appearing on the right-hand side of the equations. For example, the numerator determinants for evaluating iy, 2, and i; in Eqs. AL are 6 2 (a2)a1 -33 -12 33 2), (aa) “8 50 2 and a -9 -33 Ns=\-3 6 I (as) -3 -4 50] ‘The numerator determinants for the evaluation of v1, v2, and v5 in Eqs.A3 are 4-1 Mafie 4 3), (a0) 5s 0 4 16 3}, (aaa) 52 and 2 -1 4 N=|o 4 16) (az) 7 5 A.5 The Evaluation of a Determinant ‘The value of a determinant is found by expanding it in terms of its minors. ‘The minor of any element in a determinant is the determinant that remains after the row and column occupied by the element have been deleted. For example, the minor of the element 6 in Eq.A.7 is -33 12] so 22)" while the minor of the element 22 in Eq. A.7 is -33 -9 3 4 The cofactor of an element is its minor multiplied by the sign- controlling factor 10, where / and j denote the row and column, respectively, occupied by the element. Thus the cofactor of the element 6 in Eg. A.7 is 03-3 2 so 22 AS The Evaluation of a Determinant m7128 The Solution of Linear Simultaneous Equations and the cofactor of the element 22 is 183) -33 9 3 6 The cofactor of an element is also referred to as its signed minor. ‘The sign-controlling factor 1°? will equal +1 or ~1 depending on whether i + jis an even or odd integer. Thus the algebraic sign of a cofac- tor alternates between +1 and —1 as we move along a row or column. For a3 X 3 determinant, the plus and minus signs form the checkerboard pat- tern illustrated here: A determinant can be expanded along any row or column, Thus the first step in making an expansion is to select a row for a column j. Once a row or column has been selected, each element in that row or column is multi plied by its signed minor, or cofactor. The value of the determinant is the sum of these products. As an example, let us evaluate the determinant in. Eq. AS by expanding it along its first column, Following the rules just explained, we write the expansion as, -9 6 72 rol grcalg E]-so [2 2) aon ‘The 2 X 2 determinants in Eq. A.13 can also be expanded by minors. ‘The minor of an element in a2 X 2 determinant is a single element. It fol- lows that the expansion reduces to multiplying the upper-left element by the lower-right element and then subtracting from this product the product, of the lower-left element times the upper-right element. Using this obser- vation, we evaluate Eq.A.13 to A = 21132 — 8) + 3(—198 — 48) — 818 + 72) = 2604 — 738 — 720 = 1146. (aay) Had we elected to expand the determinant along the second row of cle- ments, we would have written = 3(-198 — 48) + (462 — 96) + 2(-84 — 72) = ~738 + 2196 — 312 = 1146, (Aa5) The numerical values of the determinants Nj, V3, and N; given by Figs. A7,A8, and A9 are N, = 1146, (A16)Ny = 2292, (aan and Ny = 3438, (aaa) It follows from Eqs. A.15 through A.18 that the solutions for j,/p,and is in Eq.A.L are (ass) and (a20) and A.6 Matrices A system of simultancous linear equations can also be solved using matrices. In what follows, we briefly review matrix notation, algebra, and terminology ‘A matrix is by definition a rectangular array of elements; thus ay ay ay Ay Aa| a a An (azn is a matrix with m rows and n columns. We describe A as being a matrix of order m by n, or m Xn, where m equals the number of rows and 7 the T Am excellent intoductory-level text in matrix application to circuit analysis is Lawrence P Heisman, Circuits, Matrices, and Linear Vector Spaces (New York: MeGraw-Hil 1963), As atvces 9730 The Solution of Linear Simultaneous Equations number of columns. We always specify the rows first and the columns sec- ond. The elements of the matrix—ay;, aj2, a;s,...—can be real numbers, complex numbers, or functions. We denote a matrix with a boldface capi tal etter, ‘The array in Eq. A.21 is frequently abbreviated by writing A= [alow (a22) where a, is the element in the ith row and the jth columa. Ifm = 1,A is called a row matrix, that is, Agfa a2 as a (a23) Im = 1, Ais called a column matrix, that is, A=] ay (ar) Ifm = n, Ais called a square matrix. For example, if m =n = 3,the square 3 by 3 matrix is ayy sy ay (425) Also note that we use brackets [] to denote a matrix, whereas we use vertical lines || to denote a determinant. tis important to know the differ- ence. A matrix isa rectangular array of clements.A determinant i a {ane- tion of a square array of elements. Thus if a matrix A is square, we can define the determinant of A. For example, it 21 Av le is then deta =|? 1] = 30-6 = 28, tA |6 1s - A.7 Matrix Algebra The equality, addition, and subtraction of matrices apply only to matrices of the same order. "TWwo matrices are equal if, and only if, their correspon- dling elements are equal. In other words, A = B if,and only if, a = ,, for all | and j. For example, the two matrices in Eqs. A.26 and A.27 are equal because ayy = bys. dia = bioy dna = boy and ayy = boy a= Ps | 4 16. (426)36-20) If A and B are of the same order, then C=A+B (az8) implies y= ay + by (nas) 4-6 , a-[i 3s (a30) and 16 10 -30) . [ee aay) then 4 20) (A.32) 20 u (a3) The equation A-B (a33) implies dy = 4 ~ by (434) For the matrices in Egs.A.30 and A.31, we wold have [-12 -16 40 Dp 28 4-19 (a5) Matrices of the same order are said to be conformable for addition and subtraction, ‘Multiplying a matrix by a scalar & is equivalent to multiplying each element by the scalar. Thus A = kB if, and only if, ay = kb, It should be noted that k may be real or complex. As an example, we will multiply the matrix D in Eq. A.35 by 5, The result is sp - [70-80 200] Ase 140-20 -95 (39) Matrix multiplication can be performed only if the number of columns in the first matrix is equal to the number of rows in the second matrix. In other words, the product AB requires the number of columns in A to equal the number of rows in B. The order of the resulting matrix will AD Natex Algebra r1732 The Solution of Linear Simultaneous Equations be the number of rows in A by the number of columns in B, Thus if C = AB, where A is of order m X p and Bis of order p X n. then C will be a matrix of order m X 7, When the number of columns in A equals the umber of rows in B, we say A is conformable to B for multiplication. ‘An clement in Cis given by the formula (aan) The formula given by Eq. A.37 is easy to use if one remembers that ‘matrix multiplication is a row-by-column operation. Hence to get the ith, jth term in C, each element in the ith row of A is multiplied by the corre- sponding element in the jth column of B, and the resulting products are summed. The following example illustrates the procedure. We are asked to, find the matrix C when a-[6 33] won Lo First we note that C-will be a 2 X 2 matrix and that each element in C will require summing three products. To find C,; we multiply the corresponding elements in row 1 of matrix A with the elements in column 1 of matrix B and then sum the products. We can visualize this multiplication and summing process by extracting the corresponding row and column from each matrix and then lining them up clement by clement. So to find Cy, we have RowlofA 6 | 3 | 2 Column lorB 4 [of 1” therefore Cy = 6X443K0F2x1= 2%, To find Cy, we visualize RowlofA 6 | 3 | 2 Column 20fB “2 [3 [=2 thus Cy = 6X 243X342 (-2) = 17 For Cy we have Row2ofA _1 | 4 4 [0and Cy H1X4+4xX OF 6X 1= 10, Finally, for C;, we have Row2ofA 1 | 4 Column 20fB 2 [3 from which Cy =1X 244X346 x (-2)=2 I: follows that c=an=[%5 Aa In general, matrix multiplication is not commutative, that is, AB # BA. As an example, consider the product BA for the matrices in Eqs.A.38 and A.39. The matrix generated by this multiplication is of order 3 X 3,and each term in the resulting matrix requires adding two products. Therefore if D = BA, we have 26 20 20 p-|3 2 18 (any 4-5 -10 Obviously, C # D, We leave you to verify the elements in Eq. Al. ‘Matrix multiplication is associative and distributive, Thus (AB)C — A(BC), (a2) A(B + C) = AB + AC, (a3) and (A + B)C = AC + BC. (aca) In Eqs.A.42,A.43, and A.44, we assume that the matrices are conformable for addition and multiplication. ‘We have already noted that matrix multiplication is not commutative. There are two other properties of multiplication in scalar algebra that do not carry over to matrix algebra First, the matrix product AB — 0 does not imply either A = 0 or B ~ 0.(Nore: A matrix is equal to zero when all its elements are zero.) For example, if then AD Natex Algebra 733134 ‘The Solution of Linear Simultaneous Equations Hence the product is zero, but neither A nor B is zero. Second, the matrix equation AB = AC does not imply B = C, For example, if 10 34 34 acts) a-Gal meni] then 34 68 -ac=[} {] tanec ‘The transpose of a matrix is formed by interchanging the rows and columns. For example, if then A? = ‘The transpose of the sum of two matrices is equal to the sum of the transposes, that is, (A + BY’ = AT + BY. (A.45) ‘The transpose of the product of two matrices is equal to the product of the transposes taken in reverse order. In other words, [ABJ = BPAY, (446) Equation A.46 can be extended to a product of any number of matri- ces. For example, [apepy’ = pTcTBTA’. (aan) If A = AT, the matrix is said to be symmetric. Only square matrices can be symmetric A.8 Identity, Adjoint, and Inverse Matrices ‘An identity matrix is a square matrix where aj, = 0 for i # j, and ay = 1 for i = j. In other words, all the elements in an identity matrix are zero, except those along the main diagonal, where they are equal to 1. Thus 1000 [op fous a fo too oo. ooif. jo e129 0001