Eigensystems 1 Eigensystems - Intro Jacob Y. Kazakia © 2005 1 Eigensystems 2 Eigensystems - Intro Jacob Y. Kazakia © 2005 2 Eigensystems 3 Eigensystems - Intro Jacob Y. Kazakia © 2005 3 Eigensystems 4 Eigensystems - Intro Jacob Y. Kazakia © 2005 4 Eigensystems 5 Eigensystems - Intro Jacob Y. Kazakia © 2005 5 Eigensystems 6 Eigensystems - Intro Jacob Y. Kazakia © 2005 6 Calculating Determinants Given a nxn matrix A as: a11 a A 21 ..... a n1 ........ a1n a22 ........ a2 n ..... ........ ..... an 2 ........ ann a12 its minor Aij is defined as the matrix obtained by eliminating the i th row and j th column. For example the minor A22 of the matrix is the (n-1)x(n-1) matrix a11 a21 A22 ..... a n1 ........ a1n a22 ........ a2 n ..... ........ ..... an 2 ........ ann a12 or a11 a A22 31 ..... a n1 ........ a1n a33 ........ a3n ..... ........ ..... an 3 ........ ann a13 k n We define the determinant by the first row expansion det(A) 1 k 1 k 1 a1k A1k here the power of -1 makes the sign alternate from positive to negative Eigensystems - Intro Jacob Y. Kazakia © 2005 7 Calculating Determinants - examples for a 2x2 matrix the determinant calculation is trivial. For example: 1 3 1 (4) 3 6 4 18 22 det 6 4 for a three by three matrix we have 5 2 4 1 3 2 3 2 1 det 2 1 3 5 2 4 4 2 3 2 3 4 3 4 2 5 (2 (12)) 2 (4 (9)) 4 (8 (3)) 5 14 2 5 4 11 70 10 44 124 Things get more difficult for a 4x4 matrix since, in the expansion we must calculate 4 , 3x3 determinants. There are other short cut ways for calculating numerical determinants. MATLAB does this effortlessly. Eigensystems - Intro Jacob Y. Kazakia © 2005 8 Systems of Differential Equations Consider the 3X3 system of first order differential equations: We write it in matrix form as: dx Ax dt 2 0 A 1 2 3 0 dx1 2 x1 3x3 dt dx2 x1 2 x2 dt dx3 3x1 2 x3 dt For each eigenvector of the matrix Ak k 1 0 A k 1 , k 2 , k 3 1 k 1 , 2 k 2 , 3 k 3 k 1 , k 2 , k 3 0 2 0 0 Eigensystems - Intro Jacob Y. Kazakia © 2005 with 3 0 2 consequently we can have 0 0 3 or equivalently: A K K D1, 2 , 3 9 Systems of Differential Equations 2 A K K D1, 2 , 3 Here K is the matrix of eigenvectors and D is a diagonal matrix. If we can find 3 linearly independent eigenvectors, then we can construct the inverse of K and hence obtain: K A K D1, 2 , 3 1 This is known as a similarity transformation and provides the means of diagonalizing a given matrix Once we know the eigenvalues and eigenvectors of the coefficient matrix, the solution of the system of differential equations can be explicitly written as: 1 t x c1e k1 c2e 2 t k 2 c3e 3 t k3 Here c1, c2, c3 are arbitrary coefficients. The derivation of this solution is shown in the next slide Eigensystems - Intro Jacob Y. Kazakia © 2005 10 Systems of Differential Equations 3 dx Ax dt In the system We then obtain: K dy dt use the transformation: A K y or dy dt xK y 1 K AK y D y c1e 1t t This produces trivially the solutions for y’s as: y c2 e 2 3t c3e c1e 1t t x K c2 e 2 The functions x are then obtained from: 3t c3e Eigensystems - Intro Jacob Y. Kazakia © 2005 11 S.D.E. 4 Complete Solution 2 0 3 A 1 2 0 3 0 2 For our matrix 2 0 3 1 2 0 3 0 2 we write the characteristic equation: 2 3 32 3 62 3 10 2 5 1 0 3 The expansion The standard form The factorization The determinant for 1 2 for 2 5 for 3 1 0 0 3 0 1 0 0 k 0 3 0 0 0 0 k1 1 0 3 3 0 0 1 3 0 k 0 3 0 0 3 3 k 2 1 3 3 0 3 0 1 3 0 k 0 3 0 3 0 3 k3 1 3 12