LECTURE 26: EIGENVALUE METHOD MINGFENG ZHAO November 16, 2015 2 Theorem 1. to ~x = A~x, 0 Let A be a 2 2 matrix, and ~y(t) = 4 then the general solution to ~x = A~x 0 y1 (t) y2 (t) 3 5 2 and ~z(t) = 4 z1 (t) z2 (t) 3 5 be two linearly independent solutions is ~x(t) = C1 ~y(t) + C2~z(t): 22 In this case, the see that ~ (t) = AX ~ (t), X 0 matrix ~ (t) = [~y(t) ~z(t)] X where 2 ~ (t) = 4 X 0 Theorem 2 (Eigenvalue Method). ~x(t) = et~v is a solution to Let is called a fundamental matrix of the system y1 (t) z1 (t) y2 (t) z2 (t) 30 5 2 := 4 be an eigenvalue of A 0 0 and ~v It's easy to 5: y2 (t) z2 (t) 0 0 3 y1 (t) z1 (t) 0 ~x = A~x. be an eigenvector corresponding to , then ~x = A~x. 0 Eigenvalue method with distinct real eigenvalues Theorem 3. Let corresponding to A 1 be a and 22 2 , matrix with two distinct real eigenvalues respectively, then the general solution to 1 ~x = A~x 0 and 2 , and ~v1 and ~v2 are eigenvectors ~v2 are eigenvectors is: ~x = C1 e1 t~v1 + C2 e2 t~v2 : Eigenvalue method with distinct complex eigenvalues Theorem 4. Let corresponding to A 1 be a and 22 2 , matrix with two distinct complex eigenvalues respectively, then 2 = 1 , ~v2 = ~v1 , 1 2 , and and the general solution to ~x = C1 Re e1 t~v1 + C2 Im e1 t~v1 : 1 and ~v1 and ~x = A~x 0 is: 2 MINGFENG ZHAO Recall the Euler's identity: ea+ib = ea cos(b) + iea sin(b): 2 Example 1. Find the general solution to the system ~x = 4 0 2 1 First, let's nd the eigenvalues of A = 4 1 1 1 3 5, 1 5~ x. 1 1 then 2 det (A 3 1 I2 ) = det 4 1 1 )2 + 1 = 2 2 + 2 5 1 = (1 3 1 = 0: Then 1 = 1 + i; and 2 = 1 i: For = 1 + i, let's solve A~v = ~v , that is, 2 4 So we get 1 1 1 1 2 4 So we have 32 54 i 1 1 i 2 4 2 That is, 4 1 i identity, we have x1 x2 x1 x2 32 54 3 2 5 = (1 + i) 4 x1 x2 3 2 5 =4 3 2 5 = x1 4 1 i 0 0 x1 x2 5: 3 5: 2 5 is an eigenvalue corresponding to 1 + i, then e(1+i)t 4 e(1+i)t 4 1 i 3 2 5 =4 e(1+i)t ie(1+i)t 3 2 5 =4 5: 3 3 2 3 et cos(t) + iet sin(t) i [et cos(t) + iet sin(t)] 3 2 5 =4 1 i 3 5 is a solution to ~x0 = A~x. By Euler's et cos(t) et sin(t) 3 2 5 + i4 et sin(t) et cos(t) 3 5: LECTURE 26: EIGENVALUE METHOD 3 Therefore, the general solution to ~x0 = A~x is: 2 3 et cos(t) ~x(t) = C1 4 et sin(t) 2 5 + C2 4 2 2 Example 2. Find the general solution to the system ~x = 4 0 2 2 First, let's nd the eigenvalues of A = 4 2 4 6 5, : 5~ x. then 2 det (A 5 et cos(t) 3 2 4 6 3 3 et sin(t) I2 ) = det 4 2 3 2 4 = (2 )(6 = 2 8 + 20 5 6 ) + 8 = 0 Then 1 = 4 + 2i; and 2 = 4 2i: For = 4 + 2i, let's solve A~v = ~v , that is, 2 2i 2 4 4 So we have 2 4 2 1 32 2 2i 2 x1 x2 54 3 2 5 = x1 4 x1 x2 3 2 5 =4 0 0 3 5: 3 1 1+i 5: 3 2 1 5 is an eigenvalue corresponding to 4 + 2i, then e(4+2i)t 4 i+i 1+i Euler's identity, we have That is, 4 2 e(4+2i)t 4 1 1+i 3 2 5 = 4 2 = 4 e(4+2i)t (1 + i)e(4+2i)t 3 2 5 =4 5 is a solution to ~x0 = A~x. By e4t cos(2t) + ie4t sin(2t) 3 (1 + i) e4t cos(2t) + ie4t sin(2t) e4t cos(2t) e4t cos(2t) 3 e4t sin(2t) 3 2 5 + i4 e4t sin(2t) e4t cos(2t) + e4t sin(2t) 5 3 5: 4 MINGFENG ZHAO Hence, the general solution to ~x0 = A~x is: 2 ~x(t) = C1 4 3 e4t cos(2t) e4t cos(2t) e4t sin(2t) 2 5 + C2 4 e4t sin(2t) e4t cos(2t) + e4t sin(2t) 3 5: Eigenvalue method with the same real eigenvalue Denition 1. Let A be a 2 2 matrix, f (t) = t2 (a + d)t + (ad bc) be the characteristic polynomial of A, and is an eigenvalue of A, that is, f () = 0, then I. The multiplicity of as a solution to f (t) is called the algebraic multiplicity of as an eigenvalue of A. II. The dimension of eigenvector space corresponding to is called the geometric multiplicity of . III. If the algebraic multiplicity and the geometric multiplicity are equal, we say that is complete IV. If the algebraic multiplicity is 2, but the geometric multiplicity is 1, we say that is a defective eigenvalue with defect 1. Let A be a 2 2 matrix with the same real eigenvalue and ~v be the eigenvector corresponding to , then ~x(t) = et~v is a solution to ~v 0 = A~x. If I2 2 A = 0, then 4 1 3 2 3 0 are two linearly independent eigenvectors corresponding to , that is, 0 1 is a complete eigenvalue of A. Hence the general solution to ~x0 = A~x is 5 and 4 5 2 ~x(t) = et 4 If I2 1 0 3 2 5 + et 4 0 1 3 5: A 6= 0, that is, is a defective eigenvalue of A, then another linearly independent solution ~y to ~x = A~x 0 has the form ~y (t) = et (t~v + w ~ ) for some vector w ~ to be determined. That is, ~y(t) = t~x(t) + et w ~ . So we get ~y (t) = ~x(t) + t~x (t) + et w ~ 0 0 = ~x(t) + tA~x(t) + et w ~ = A~y (t) = tA~x(t) + et Aw: ~ Then ~x(t) + et w ~ = et Aw ~ , that is, et~v + et w ~ = et Aw ~ , so w ~ is a solution to ~v + w ~ = Aw ~ , that is, (A I2 )w ~ = ~v : LECTURE 26: EIGENVALUE METHOD In this case, we have (A I2 )2 w ~ = (A 5 I2 )~v = 0, we call w ~ a generalized eigenvector of A. Problems you can do: Lebl's Book [2]: All Exercise on Page 108, Page 109, Page 127 and Page 128. Braun's Book [1]: All exercises on Page 344, Page 345, Page 352, Page 353, Page 354 and Page 355. Read all materials in Section 3.9 and Section 3.10. References [1] Martin Braun. [2] Jiri Lebl. Dierential Equations and Their Applications: An Introduction to Applied Mathematics. Springer, 1992. Notes on Diy Qs: Dierential Equations for Engineers. Createspace, 2014. 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