矩阵的inverse转换.ppt

The inverse of a matrix .Definition .Purpose .Properties .Calculation method (How) .Calculation method (Why) .Numerical note Definition A square matrix B is called the inverse of a square matrix A, if A.B=B.A=I. We use the notation: B=A-1. An inverse of A exists if and only if. A is square. The columns of A are linearly independent Purpose Suppose we have a linear transformation:T: Rn? Rn, having matrix A: T(x)=A.x=y. If A has an inverse we have: A-1.y=A-1(A.x) = (A-1.A)x = I.x = x A-1 is the matrix belonging to the inverse transformation T-1. A has an inverse implies that T is a bijection. A is square Properties A-1 is unique A.B=I has only 1 solution B if A is invertible! =  A.D=C has only 1 solution D. (A-1)-1=A (AB)-1=B-1A-1 (AT)-1=(A-1)T Calculation method (How (1)) Sweep the Augmented matrix [A I] to [I A-1]. A is invertible, because the columns are linearly independent. A I I A-1 Calculation method (How(2)) Only for a 2x2 matrix: If det(A)=0, the determinant of A=0, then A is not invertible! Calculation method (How(3)) Knowledge of the transform T, which A represents Horizontal shear, 1.9 Table 3 Calculation method (Why) We know: . A-1y=x and Ax=y. . A-1e1= column 1 of A-1=x, solve Ax=e1 . A-1e2= column 2 of A-1=x, solve Ax=e2 . Solving both equations separately can be done by sweeping the augmented matrices [A e1] and [A e2] . Combine the two actions in one go!Sweep: [A e1 e2]=[A I]. Numerical note Solving Ax=y for x is mostly not performed by calculating the inverse of A! Reasons: . Number of operations is big . Numerical (in)stability of the process Instead decomposition methods are preferred: . LU-decomposition (Ch 2.5) . QR-decomposition (Ch 6.3)