Math 317: Linear Algebra Homework 11 (Extra Credit) Due: December 11, 2015 The following problems are for additional practice and are not to be turned in: (All problems come from Linear Algebra: A Geometric Approach, 2nd Edition by ShifrinAdams.) Turn in the following problems. 1. Section 6.1, Problem 1c 10 −6 We want to find the eigenvalues and eigenvectors for A = . To begin, 18 −11 we solve the characteristic equation det(A − λI) = 0 =⇒ 10 − λ −6 det = 0 =⇒ (10 − λ)(−11 − λ) − (18)(−6) = 0 =⇒ 18 −11 − λ (λ + 2)(λ − 1) = 0 =⇒ λ1 = −2, λ2 = 1. To find the corresponding eigenvectors, we find bases for N (A − (−2)I) and N (A − 1I). To this extent, we obtain 12 −6 2 −1 A + 2I = → =⇒ 2x1 − x2 = 0, x2 = x2 =⇒ x1 = 18 −9 0 0 1/2 (1/2)x2 , x2 = x2 =⇒ v1 = is an eigenvector for λ1 . The second eigenvec1 tor comes from N (A − I) and so 9 −6 3 −2 2/3 A−I = → =⇒ x1 = (2/3)x2 , x2 = x2 =⇒ v2 = is 18 −12 0 0 1 the second eigenvector that corresponds to λ2 . 2. Section 6.1, Problem 2 Proof : Recall that A is singular if and only if det(A) = 0 which is true if and only det(A − 0I) = 0 which is true if and only 0 is an eigenvalue for A. (Alternatively, you could have argued that A is singular if and only Ax = 0 has a nontrivial solution which implies that Ax = 0x for some nonzero x which implies that 0 is an eigenvalue with the corresponding eigenvector x.) 3. Section 6.1, Problem 4 Recall that if A is the projection matrix that cooresponds to a subspace V ⊂ Rn then Av = v if v ∈ V and Av = 0 if v ∈ V ⊥ . Since Rn = V + V ⊥ we know that all vectors in Rn will either be in V or V ⊥ . Thus Av = 1v or Av = 0v. This tells us that λ1 = 1 and λ2 = 0. The eigenvectors that correspond to λ1 = 1 come from the nonzero vectors in V = C(A). The eigenvectors that correspond to λ2 = 0 come from the nonzero vectors in V ⊥ = N (AT ). Similarly, we find the eigenvalues and eigenvectors for a reflection matrix in the same manner. Noting that R(x) = projV (x) − projV ⊥ (x) for all x ∈ Rn , we have that Ax = x if x ∈ V and Ax = −x if x ∈ V ⊥ . Thus the eigenvalues of A are 1, −1 and the eigenvectors that correspond to 1 are given by those nonzero vectors in V and 1 Math 317: Linear Algebra Homework 11 (Extra Credit) Due: December 11, 2015 the eigenvectors that correspond to −1 are given by those nonzero vectors in V ⊥ 4. Section 6.1, Problem 7 (a) Proof : Suppose that x is an eigenvector for A that corresponds to an eigenvalue λ. Then Ax = λx. Multiply both sides by A to obtain A2 x = λ(Ax) = λ(λx) = λ2 x. If we continue this process by multiplying A n − 1 times to A, we obtain An x = λn x which implies that λn is an eigenvalue of An with the corresponding eigenvector x. (b) Proof : We claim that x is an eigenvector of A + I. This is true since (A + I)x = Ax + Ix = λx + x = (λ + 1) x which implies that x is an eigenvector of A + I with the corresponding eigenvalue λ + 1. (c) Proof : This claim is true. Let x be an eigenvector of B with the corresponding eigenvalue µ. Recalling that x is an eigenvector for A that corresponds to an eigenvalue λ, we have that (A+B)x = Ax+Bx = λx+µx = (λ+µ)x. 10 −6 3 1 (d) This claim is false. To see why, let A = and B = . An 18 −11 −3 7 eigenvalue for A is given by λ =1. An eigenvalue forB is given by λ = 4. 13 −5 8 −5 We see that A + B = . (A + B) − 5I = . Note that 15 −4 15 −9 det((A + B) − 5I)) = 8(−9) − (15)(−5) 6= 0 which means that (A + B) − 5I is nonsingular. Hence 5 = 1 + 4 is not an eigenvalue for A + B. 5. Section 6.1, Problem 10 (a) Proof : It is sufficient to show that A and AT have the same characteristic equation. Recalling that det(A) = det(AT ) we have that det(A − λI) = det((A − λI)T ) = det(AT − (λI)T ) = det(AT − λI). So A and AT have the same characteristic equation and hence they have the same eigenvalues. 1 1 (b) This claim is false. To see why, let A = . Then A has eigenvalues 0 0 1 1 0, 1 with associated eigenvectors v1 = , v2 = . However, AT has −1 0 0 1 eigenvectors w1 = and w2 = . 1 1 6. Section 6.2, Problem 2 (a) Proof : Suppose that A is an n × n matrix with n distinct eigenvalues. Then the algebraic multiplicity of each eigenvalue is 1. That is, alg mult(λi ) = 1 for all i = 1, . . . , n. Recalling that alg mult(λi ) ≥ geo mult(λi ) and that geo mult(λi ) > 0, we have that geo mult(λi ) = 1 for each i = 1, . . . n. Thus, alg mult(λi ) = geo mult(λi ) for each i and hence A is diagonalizable. 2 Math 317: Linear Algebra Homework 11 (Extra Credit) Due: December 11, 2015 1 0 (b) This claim is false. To see why let A = which is a diagonal matrix 0 1 2 1 and hence is diagonalizable. Let B = . We have shown that B is not 0 2 diagonalizable. Now, AB = BA but B is not diagonalizable, even though A is diagonalizable. (c) Proof : To show that A and B have the same eigenvalues, it is sufficient to show that A and B have the same characteristic equation. Using the fact that A = P −1 BP we have that det(A − λI) = det(P −1 BP − λI) = det(P −1 BP − P −1 λIP ) = det(P −1 (B − λI)P ) = det(P −1 ) det(B − λI) det(P ) = det(B − λI). Thus A and B have the same characteristic polynomial and hence A and B have the same eigenvalues. 2 0 2 1 (d) This is not true. To see why, let A = and B = . Then A and 0 2 0 2 B have the same eigenvalues but are not similar to each other as we have shown in a previous homework assignment. 1 1 (e) This statement is false. Let A = . Then A2 = −I. −2 −1 (f) Proof : Suppose that A and B are diagonalizable with the same eigenvalues and algebraic multiplicities. Then A and B are similar to the same diagonal matrix D. That is, there is an invertible P such that A = P DP −1 and an invertible matrix Q such that B = QDQ−1 . Thus Q−1 BQ = D and hence A = P (Q−1 BQ)P −1 . This implies that A = (QP −1 )−1 B(QP −1 ), and hence A is similar to B. 7. Section 6.2, Problem 11 (a) Proof : Suppose that A is an n × n matrix with the property that A2 = A. Furthermore, suppose that λ is an eigenvalue of A. Then Ax = λx. Multiplying both sides by A we obtain A2 x = A(Ax) = A(λx) = λ2 x = Ax = λx, where we have used the fact that A2 = A. Thus λ2 x = λx =⇒ (λ2 − λ)x = 0 =⇒ λ2 − λ = 0 =⇒ λ(λ − 1) = 0 =⇒ λ = 0 or λ = 1. (b) By definition, the eigenvectors associated with λ = 0 is given by N (A). The eigenvectors associated with λ = 1 is given by N (A − I) =⇒ Ax = 1x =⇒ x ∈ C(A) by exercise 3.2.13. Since C(A) ∩ N (A) = {0} and C(A) + N (A) = Rn , then the eigenvectors from a linearly independent spanning set for Rn . That is, the eigenvectors of A form a basis for Rn and hence A is diagonalizable. 8. Problem 6.3, Problem 1 3 Math 317: Linear Algebra Homework 11 (Extra Credit) Due: December 11, 2015 We wish to calculate Ak for any k ≥ 1. To begin, we find the eigenvalues 2−λ 5 for A by solving det(A − λI) = 0 =⇒ det = 0 =⇒ 1 −2 − λ (2 − λ)(−2 − λ) − (5) = 0 =⇒ (λ + 3)(λ − 3) = 0 =⇒ λ1 = −3, λ2 = 3. Next 5 5 we find the corresponding eigenvectors. For λ = −3, A − (−3I) = =⇒ 1 1 1 −1 5 x1 = −x2 , x2 = x2 =⇒ v1 = . Similarly, A − 3I = =⇒ x1 = −1 1 −5 5 5x2 , x2 = x2 =⇒ v2 = . Recalling that Ak = P Dk P −1 where the columns 1 of P contain the eigenvectors of A, we have that k −1 5 1 3 0 5 1 A = 1 −1 0 (−3)k 1 −1 k 4