* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download MATH321 – HOMEWORK SOLUTIONS HOMEWORK #5 Section 2.4
Eigenvalues and eigenvectors wikipedia , lookup
Matrix calculus wikipedia , lookup
Non-negative matrix factorization wikipedia , lookup
Bra–ket notation wikipedia , lookup
System of linear equations wikipedia , lookup
Cayley–Hamilton theorem wikipedia , lookup
Linear algebra wikipedia , lookup
MATH321 – HOMEWORK SOLUTIONS HOMEWORK #5 Section 2.4: Problems 1, 2, 3, 4, 5, 6, 7, 13, 14, 22 Section 2.5: Problems 1, 2(a)(b), 3(a), 4, 6, 7, 9, 10 Krzysztof Galicki Problem 2.4.1 (See Answers to Selected Exercises). Problem 2.4.2 (a) T : R2 −→R3 cannot be invertible. However, N (T ) = {(0, 0)} is trivial so that T can be inverted on R(T ) which is a plane in R3 . (b) T : R2 −→R3 cannot be invertible. N (T ) = {(0, 0)} is trivial so that T can be inverted on R(T ) which is a plane in R3 . (c) Here N (T ) = {(0, 0, 0)} and it follows that R(T ) = R3 by nullity-rank equation. Hence, T is invertible. (d) Here N (T ) is the space of constant polynomials f (x) = c. Hence T is not one-to-one and cannot be inverted. 0 0 (e) Here N (T ) consists of matrices of the form . Hence T is not oon-to-one c −c and cannot be inverted. (f) Here N (T ) is trivial so it follows that R(T ) = M2×2 (R) by nullity-rank equation. Hence, T is both one-to-one and onto. T is invertible. Problem 2.4.3 (a) F 3 and P3 (F ) are of dimension 3 and 4, respectively so thay cannot be isomorphic. (b) F 4 and P3 (F ) are isomorphic. One could take T : P3 (F )−→F 4 to be T (a + bx + cx2 + dx3 ) = (a, b, c, d). Clearly, T is invertible. (c) Take T a b c d = a + bx + cx2 + dx3 . This map is linear and invertible. (d) V is 3-dimensional so it cannot be isomorphic to R4 . Problem 2.4.4 Let A, B ∈ Mn×n (F ) be invertible with inverses A−1 and B−1 . Consider the product C = B−1 A−1 . We have C(AB) = B−1 A−1 AB = B−1 B = In and, similarly, λ(AB)C = ABB−1 A−1 = AA−1 = In . It follows that AB is invertible with inverse C. Problem 2.4.5 Let A ∈ Mn×n (F ) be invertible with inverse A−1 . Consider C = (A−1 )t . We have CAt = (A−1 )t At = (AA−1 )t = Itn = In . Similarly At C = At (A−1 )t = (A−1 A)t = Itn = In . Note, that we use (AB)t = Bt At which is true for any matrices. Verify this. Problem 2.4.6 Let A ∈ Mn×n (F ) be invertible with inverse A−1 . If AB = On we can multiply this equation by A−1 from the left A−1 AB = A−1 On The RHS equals On because multiplying any matrix by zero matrix gives zero. On the other hand A−1 AB = B which gives B = On . Problem 2.4.7 (a) This follows from the proof in Problem 2.4.6. Assuming otherwise we would get A = On which is contradiction with A invertible. (b) No. Again, this follows from Problem 2.4.6. Problem 2.4.13 Recall that an equivalence relation ∼ is a relation which is reflexive, symmetric, and transitive. Isomorphism between vector spaces over a field F is an equivalence relation because: 1. It is reflexive. IV is an isomorphism between V and V . 2. It is symmetric. If V is isomorphic to W then there exists a linear map T : V −→W which is invertible. But then T −1 : W −→V is also invertible so that W is isomorphic to V. 3. It is transitive. If V is isomorphic to W there exists a linear map T : V −→W which is one-to-one and onto. If W is isomorphic to Z there exists a linear map U : W −→Z which is one-to-one and onto. But then the composition U ◦ T : V −→Z is also one-to-one and onto (see Problem 2.3.12(c)). Problem 2.4.14 Note that V is the space of upper triangular two by two matrices. We can define x y T = (x, y, z). 0 z Clearly, T is linear, N (T ) is trivial so that we must have R(T ) = F 3 . Hence, T is an isomorphism. Problem 2.4.22 Choose any distinct vector (c0 , c1 , . . . , cn ) ∈ F n+1 such that all coordinates are distinct. Let T : Pn (F ) → F n+1 be defined as T (f ) = (f (c0 ), f (c1 ), . . . , f (cn )). We must show that T is linear. This follows from the definition as T (f + ag) = (f (c0 ) + ag(c0 ), . . . , f (cn ) + ag(cn )) = = (f (c0 ), f (c1 ), . . . , f (cn )) + a(g(c0 ), g(c1 ), . . . , g(cn )) = T (f ) + aT (g). Consider the null space N (T ). The only polynomial g(x) of degree n which vanishes at n + 1 distinct points Pn is the zero polynomial. This follows from the Lagrange interpolation formula g(x) = i=0 g(ci )fi (x) on page 52 of your text. Hence, N (T ) is trivial and T is one-to-one. By the rank-nulltity equation R(T ) = F n+1 so that T is also onto. T is an isomorphism. Problem 2.5.1 (See Answers to Selected Exercises). Problem 2.5.2 (a) We have IR2 (a1 , a2 ) = (a1 , a2 ) = a1 e1 + a2 e2 , IR2 (b1 , b2 ) = (b1 , b2 ) = b1 e1 + b2 e2 so that, by definition, Q= a1 a2 b1 b2 . (b) We have IR2 (5, 0) = (5, 0) = (−1, 3) + 3(2, −1) IR2 (0, 10) = (0, 10) = 4(−1, 3) + 2(2, −1), so that, by definition, Q= 4 2 1 3 . Problem 2.5.3 (a) We have IP2 (R) (a2 x2 + a1 x + a0 ) = a2 x2 + a1 x + a0 , IP2 (R) (b2 x2 + b1 x + b0 ) = b2 x2 + b1 x + b0 , IP2 (R) (c2 x2 + c1 x + c0 ) = c2 x2 + c1 x + c0 , so that, by definition, a2 Q = a1 a0 b2 b1 b0 c2 c1 . c0 Problem 2.5.4 First observe (see Problem 2.5.2(a)) that the matrix 1 1 Q= 1 2 is the coordinate change matrix between β 0 and the standard basis β = {e1 , e2 }. Now, relative to β a 2 1 a T = b 1 −3 b so that [T ]β = 2 1 1 −3 . We get [T ]β 0 = Q −1 [T ]β Q = 2 −1 −1 1 2 1 1 −3 1 1 1 2 = 8 −5 13 −9 . Problem 2.5.6 (a) First observe (see Problem 2.5.2(a)) that the matrix 1 1 Q= 1 2 is the coordinate change matrix between β and the standard basis {e1 , e2 }. Hence, [LA ]β = Q −1 AQ = 2 −1 −1 1 1 3 1 1 1 1 1 2 = 6 11 −2 −4 . Problem 2.5.7 Remember, every transformation is determined by what it does to a basis. (a) Consider the standard basis {(1, 0), (0, 1)}. If T is a reflection through the y = mx line then we must have T (1, m) = T (1, 0) + mT (0, 1) = (1, m), T (m, −1) = mT (1, 0) − T (0, 1) = (−m, 1). This follows form the fact that (1, m) is on L and (−m, 1) is perpendicular to L. In particular, (1, m) is its own image and the image of (−m, 1) is (m, −1) One can easily solve these equations to get explicit formulas for T (1, 0), T (0, 1) T (1, 0) = T (0, 1) = 1 − m2 2m , 2 1 + m 1 + m2 m2 − 1 2m , . 1 + m2 1 + m2 Thus 2m 1 − m2 , T (x, y) = xT (1, 0) + yT (0, 1) = x 2 1 + m 1 + m2 2m m2 − 1 +y , . 1 + m2 1 + m2 (a) Consider the standard basis {(1, 0), (0, 1)}. If T is an orthogonal projection on the y = mx line then we must have T (1, m) = T (1, 0) + mT (0, 1) = (1, m), T (m, −1) = mT (1, 0) − T (0, 1) = (0, 0). One can easily solve these equations to get explicit formulas for T (1, 0), T (0, 1) T (1, 0) = 1 m , 2 1 + m 1 + m2 T (0, 1) = mT (1, 0) = m m2 , . 1 + m2 1 + m2 Thus m 1 . , T (x, y) = xT (1, 0) + yT (0, 1) = (x + my) 1 + m2 1 + m2 Problem 2.5.9 Recall that A ∼ B if there exists C such that A = C−1 BC. Now, A ∼ A as we can take C = In so the relation is reflexive. It is symmetric as if A = C−1 BC then B = D−1 AD, where D = C−1 . To show transitivity we assume X ∼ Y ≡ X = C−1 YC, Y ∼ Z ≡ Y = D−1 ZD. Then X = C−1 YC = C−1 D−1 ZDC = (DC)−1 Z(DC) ≡ X ∼ Z. Problem 2.5.10 If A ∼ B there exists C such that A = C−1 BC. But then Tr(A) = Tr(C−1 BC) = Tr(BC−1 C) = Tr(B).