LECTURE 25: EIGENVALUE METHOD AND MULTIPLE EIGENVALUE MINGFENG ZHAO November 03, 2014 Theorem 1. Let A be a 2 × 2 matrix, ~y (t) and ~z(t) be two linearly independent solutions to ~x0 = A~x, then the general solution to ~x0 = A~x is ~x(t) = C1 ~y (t) + C2 ~z(t). ~ In this case, the 2 × 2 matrix X(t) = [~y (t) ~z(t)] is called a fundamental matrix of the system ~x0 = A~x. It’s easy to ~ 0 (t) = AX(t). ~ see that X Theorem 2 (Eigenvalue Method). Let λ be an eigenvalue of A and ~v be an eigenvector corresponding to λ, then ~x(t) = eλt~v is a solution to ~x0 = A~x. Eigenvalue method with distinct eigenvalues Theorem 3. Let A be a 2×2 matrix with two distinct eigenvalues λ1 and λ2 , and ~v1 and ~v2 are eigenvectors corresponding to λ1 and λ2 , respectively. I. If λ1 and λ2 are two different real eigenvalues, then the general solution to ~x0 = A~x is: ~x = C1 eλ1 t~v1 + C2 eλ2 t~v2 . II. If λ1 and λ2 are two different complex eigenvalues, then λ2 = λ1 , ~v2 = ~v1 , and the general solution to ~x0 = A~x is: ~x = C1 Re eλ1 t~v1 + C2 Im eλ1 t~v1 . Example 1. Solve ~x0 = 2 2 1 3 ~x. 1 2 MINGFENG ZHAO First, let’s find eigenvalues of A := 2 2 1 3 det (λI2 − A) = det , we need to solve λ−2 −2 −1 λ−3 = (λ − 2)(λ − 3) − 2 = λ2 − 5λ + 4 = 0. Then λ1 = 1, and λ2 = 4. For λ1 = 1, let’s find the eigenvalue corresponding to λ1 = 1, we need to solve A~x = ~x, that is, −1 −2 x1 0 = . −1 −2 x2 0 Then x1 = x2 x2 −2 . That is, 1 −2 is an eigenvalue corresponding to λ1 = 1. 1 For λ2 = 4, let’s find the eigenvalue corresponding to λ2 = 4, we need to solve A~x = 4~x, that is, 2 −2 x1 0 = . −1 1 x2 0 Then x1 x2 = x1 1 , that is, 1 1 is an eigenvalue corresponding to λ2 = 4. 1 0 Therefore, the general solution to ~x = 2 2 1 3 ~x is: ~x(t) = C1 et −2 1 + C2 e4t 1 1 . LECTURE 25: EIGENVALUE METHOD AND MULTIPLE EIGENVALUE 2 2 −4 6 Example 2. Solve the initial value problem ~x0 = First, let’s find the eigenvalues of A = det (λI2 − A) = det 2 2 −4 6 1 ~x and ~x(0) = 3 . 1 , then λ−2 −2 4 λ−6 = (λ − 2)(λ − 6) + 8 = λ2 − 8λ + 20 = 0. Then λ1 = 4 + 2i, and λ2 = 4 − 2i. For λ = 4 + 2i, let’s solve A~v = λ~v , that is, −2 2 + 2i −2 + 2i 4 x1 = x2 0 . 0 So we have x1 = x1 x2 That is, 1 1 . 1+i is an eigenvalue corresponding to 4 + 2i, then e(4+2i)t i+i Euler’s identity, we have 1 is a solution to ~x0 = A~x. By 1+i e(4+2i)t 1 1+i e4t cos(2t) + ie4t sin(2t) = = (1 + i)e(4+2i)t (1 + i) e4t cos(2t) + ie4t sin(2t) e4t cos(2t) e4t sin(2t) + i . = e4t cos(2t) − e4t sin(2t) e4t cos(2t) + e4t sin(2t) e(4+2i)t Hence, the general solution to ~x0 = A~x is: ~x(t) = C1 e4t cos(2t) 4t 4t e cos(2t) − e sin(2t) + C2 e4t sin(2t) 4t 4t e cos(2t) + e sin(2t) . 4 MINGFENG ZHAO On the other hand, since ~x(0) = 1 , then 1 C1 1 0 + C2 1 Then C1 = 0 and C2 = 0. Therefore, the solution to ~x0 = 1 = 1 ~x(t) = . 1 2 2 −4 6 ~x and ~x(0) = 1 is: 1 e4t cos(2t) e4t cos(2t) − e4t sin(2t) . Eigenvalue method with the same real eigenvalue Let A be a 2 × 2 matrix with the same real eigenvalue λ and ~v be the eigenvector corresponding to λ, then ~x(t) = eλt~v is a solution to ~v 0 = A~x. If λI2 − A = 0, then the general solution to ~x0 = A~x is 1 0 + eλt . ~x(t) = eλt 0 1 If λI2 − A 6= 0, then another linearly independent solution ~y to ~x0 = A~x has the form ~y (t) = eλt (t~v + w) ~ for some vector w ~ to be determined. That is, ~y (t) = t~x(t) + eλt w. ~ So we get ~y 0 (t) = ~x(t) + t~x0 (t) + λeλt w ~ = ~x(t) + tA~x(t) + λeλt w ~ = A~y (t) = tA~x(t) + eλt Aw. ~ Then ~x(t) + λeλt w ~ = eλt Aw, ~ that is, eλt~v + λeλt w ~ = eλt Aw, ~ so w ~ is a solution to ~v + λw ~ = Aw, ~ that is, (λI2 − A)w ~ = −~v . x0 = 3x1 + x2 1 . Example 3. Find the general solution to x0 = 3x 2 2 x0 = 3x1 + x2 3 1 The coefficient matrix of the system is A = x0 = 3x 0 2 2 det (λI2 − A) = det λ−3 −1 0 λ−3 1 . First, let’s find eigenvalues of A, then 3 = (λ − 3)2 = 0. LECTURE 25: EIGENVALUE METHOD AND MULTIPLE EIGENVALUE 5 Then λ1 = λ2 = 3. Now let’s solve 3 1 0 3 0 −1 x1 x1 = 3 x2 . x2 Then So x1 = x1 x2 1 x1 0 0 = x2 0 . 0 . Hence we know that λ = 3 is the only one eigenvalue for A, and ~v = 0 3t 1 is the eigenvector. 0 0 Then ~y (t) = e ~v is a solution to ~x = A~x. Let ~x = e3t (t~v + w) ~ = t~y (t) + e3t w ~ be another linearly independent solution to ~x0 = A~x, then ~x0 = ~y (t) + t~y 0 (t) + 3e3t w ~ = A(t~y (t) + e3t w) ~ = tA~y (t) + e3t Aw. ~ Since ~y (t) is a solution to ~x0 = A~x, then ~y (t) + 3e3t w ~ = e3t Aw. ~ Since ~y (t) = e3t~v , then e3t~v + 3e3t w ~ = e3t Aw, ~ that is, ~v + 3w ~ = Aw. ~ Hence, (3I2 − A)w ~ = ~v , that is, 0 −1 w 1 1 = 0 0 w2 0 So we can take w1 = 0 and w2 = −1, that is, w ~ = e3t (t~v + w) ~ = e3t t 0 −1 1 . So another solution is + 0 0 −1 = e3t t −1 . Therefore, the genera solution to ~x0 = A~x is: 3t ~x = C1 e 1 0 3t + C2 e t −1 . Department of Mathematics, The University of British Columbia, Room 121, 1984 Mathematics Road, Vancouver, B.C. Canada V6T 1Z2 E-mail address: mingfeng@math.ubc.ca