MATRICES. a m,1 a m,n A =

Transcription

1 MATRICES Matrices are rectangular arrays of real or complex numbers With them, we define arithmetic operations that are generalizations of those for real and complex numbers The general form a matrix of order m n is A = a 1,1 a 1,n a m,1 a m,n We say it has order m n Matrices that consist of a single column are called column vectors, and those consisting of a single row are called row vectors In both cases, they will have properties identical to the geometric vectors studied earlier in mulitvariable calculus We assume that most of you have seen this material previously in a course named linear algebra, matrix algebra, or something similar Section 62 in the text is intended as both a quick introduction and review of this material

2 MATRIX ADDITION Let A = [a i,j and B = [b i,j be matrices of order m n Then C = A + B is another matrix of order m n, with c ij = a i,j + b i,j EXAMPLE =

3 MULTIPLICATION BY A CONSTANT a 1,1 a 1,n c a m,1 a m,n = ca 1,1 ca 1,n ca m,1 ca m,n EXAMPLE [ a b ( 1) c d = [ a b = c d

4 THE ZERO MATRIX 0 Define the zero matrix of order m n as the matrix of that order having all zero entries It is sometimes written as 0 m n, but more commonly as simply 0 Then for any matrix A of order m n, A + 0 = 0 + A = A The zero matrix 0 m n acts in the same role as does the number zero when doing arithmetic with real and complex numbers EXAMPLE [ [ = [

5 We denote by A the solution of the equation A + B = 0 It is the matrix obtained by taking the negative of all of the entries in A For example, [ [ [ a b a b = c d c d 0 0 [ [ a b a b = c d c d [ a1,1 a 1,2 = a 2,1 a 2,2 [ a b = ( 1) c d [ a1,1 a 1,2 a 2,1 a 2,2

6 MATRIX MULTIPLICATION Let A = [a i,j have order m n and B = [b i,j have order n p Then C = AB is a matrix of order m p and c i,j = A i, B,j = a i,1 b 1,j + a i,2 b 2,j + + a i,n b n,j or equivalently c i,j = [a i,1 a i,2 a i,n b 1,j b 2,j b n,j = a i,1 b 1,j + a i,2 b 2,j + + a i,n b n,j

7 EXAMPLES [ [ a 1,1 a 1,n a n,1 a n,n x 1 x n Thus we write the linear system = = = [ a 1,1 x a 1,n x n a n,1 x a n,n x n as Ax = b a 1,1 x a 1,n x n = b 1 a n,1 x a n,n x n = b n

8 THE IDENTITY MATRIX I For a given integer n 1, Define I n to be the matrix of order n n with 1 s in all diagonal positions and zeros elsewhere: I n = 0 1 More commonly it is denoted by simply I Let A be a matrix of order m n Then AI n = A, I m A = A The identity matrix I acts in the same role as does the number 1 when doing arithmetic with real and complex numbers

9 THE MATRIX INVERSE Let A be a matrix of order n n for some n 1 We say a matrix B is an inverse for A if AB = BA = I It can be shown that if an inverse exists for A, then it is unique EXAMPLES If ad bc 0, then [ 1 [ a b 1 d = c d ad bc c [ [ = = b a

10 Recall the earlier theorem on the solution of linear systems Ax = b with A a square matrix Theorem The following are equivalent statements 1 For each b, there is exactly one solution x 2 For each b, there is a solution x 3 For each b, there is at most one solution x 4 The homogeneous system Ax = 0 has only the solution x = 0 5 det (A) 0 6 A 1 exists

11 EXAMPLE det Therefore, the linear system x 1 x 2 x 3 = 0 = is not always solvable, the coefficient matrix does not have an inverse, and the homogeneous system Ax = 0 has a solution other than the zero vector, namely = b 1 b 2 b

12 The arithmetic properties of matrix addition and multiplication are listed on page 248, and some of them require some work to show For example, consider showing the distributive law for matrix multiplication, (AB) C = A (BC) with A, B, C matrices of respective orders m n, n p, and p q, respectively Writing this out, we want to show p (AB) i,k C k,l = k=1 n A i,j (BC) j,l j=1 for 1 i m, 1 l q With new situations, we often use notation to suggest what should be true But this is done only after deciding what actually is true Read carefully the properties given in the text on page 248

13 PARTITIONED MATRICES Matrices can be built up from smaller matrices; or conversely, we can decompose a large matrix into a matrix of smaller matrices For example, consider A = B = [ [ B c = d e [ 0 c = d = [ 0 1 e = 5 1 MATLAB allows you to build up larger matrices out of smaller matrices in exactly this manner; and smaller matrices can be defined as portions of larger matrices We will often write an n n square matrix in terms of its columns: A = [A,1,, A,n For the n n identity matrix I, we write I = [e 1,, e n, with e j denoting a column vector with a 1 in position j and zeros elsewhere

14 ARITHMETIC OF PARTITIONED MATRICES As with matrices, we can do addition and multiplication with partitioned matrices provided the individual constituent parts have the proper orders For example, let A, B, C, D be n n matrices Then [ [ [ I A I C I + AD C + A = B I D I B + D I + BC Let A be n n and x be a column vector of length n Then Ax = [A,1,, A,n Compare this to a 1,1 a 1,n a n,1 a n,n x 1 x n x 1 x n = x 1 A,1 + + x n A,n = a 1,1 x a 1,n x n a n,1 x a n,n x n

15 PARTITIONED MATRICES IN MATLAB In MATLAB, matrices can be constructed using smaller matrices For example, let A = [1, 2; 3, 4; x = [5, 6; y = [7, 8 ; Then forms the matrix B = [A, y; x, 9; B =

16 ELEMENTARY ROW OPERATIONS As preparation for the discussion of Gaussian Elimination in Section 63, we introduce three elementary row operations on general rectangular matrices They are: i) Interchange of two rows ii) Multiplication of a row by a nonzero scalar iii) Addition of a nonzero multiple of one row to another row Consider the rectangular matrix A = We add row 2 times ( 1) to row 1, and then add row 3 times ( 1) to row 2 to obtain the matrix:

17 Add row 2 times ( 1) to row 1, and to row 3 as well: Add row 1 times ( 2) to row 3: Interchange row 1 and row 2: Finally, we multiply row 3 by 1/3: /3

18 /3 This is obtained from A using elementary row operations A reverse sequence of operations of the same type converts this result back to A