HOMEWORK 2 SOLUTIONS Levandosky, Linear Algebra 2.11 There are many possible parameterizations. Following the notation in the 3 text, we can first let x0 = , and obtain a direction vector using the 1 3−2 vector whose tail is at (2, −3) and head is at (3, 1): v = = 1 − (−3) 1 . Thus, the line can be written parametrically as 4 3 1 L= +t |t∈R . 1 4 1 2.13 As above, one such parameterization is given by x0 = 2 and v = 3 1 − (−2) 3 2 − 1 = 1 ; i.e. 3−2 1 3 1 L = 2 + t 1 | t ∈ R . 3 1 1 2.15 We can first let x0 = 2 . To obtain two direction vectors, we can use 3 the vectors whose tails are at (1, 2, 3) and whose heads are at (2, 3, 4)and 2−1 1 2−1 (2, 1, 5). This gives v1 = 3 − 2 = 1 and v2 = 1 − 2 = 4−3 1 5−3 1 −1 . Thus, the plane can be written paramterically as 2 1 1 1 P = 2 + s 1 + t −1 | s, t ∈ R . 3 1 2 2 3 2.17 We can first let x0 = 4 . To obtain three direction vectors, we can use 5 the vectors whose tails are at (2, 3, 4, 5) and whose heads are at (1, 1, 1, 1), 1 2 HOMEWORK 2 SOLUTIONS 1−2 1−3 (0, 1, 0, 1), and (−1, −2, 3, 1). This gives v1 = 1−4 = 1−5 0−2 −2 −1 − 2 −3 1−3 −2 −2 − 3 −5 v2 = 0 − 4 = −4 , and v3 = 3 − 4 = −1 . 1−5 −4 1−5 −4 the hyperplane can be written paramterically as 2 −1 −2 −3 −2 −2 −5 3 P = +r + s + t | r, s, t ∈ R . −4 −1 4 −3 5 −4 −4 −4 −1 −2 , −3 −4 Thus, 2 1 4.1 (a) a · b = 1 · −2 = 2 · 1 + 1 · −2 + 1 · 0 = 0. 1 0 (b) 2 1 4 2 5 a · (b + c) = 1 · −2 + 2 = 1 · 0 1 0 −2 1 −2 = 2 · 5 + 1 · 0 + 1 · −2 = 8. (c) x · y = 2 3 −3 · = 2 · −3 + 3 · 2 = 0. 2 (d) 2 −3 = 5 kx − yk = − 3 2 1 p = 52 + 12 √ = 26. 1 p √ 2 2 4.4 Note that −3 = 1 + (−3) = 10. So, to get a vector pointing in the same direction, but with length 3, we simply scale the given vector " √3 # 1 10 by √310 . So, the desired vector is √310 = √−9 . −3 10 4.10 Suppose that c1 u + c2 v + c3 w = 0 for some ci ∈ R. We must show that c1 = c2 = c3 = 0. To show this, first dot product both sides of the above equation by u. This gives c1 (u · u) + c2 (u · v) + c3 (u · w) = u · 0. HOMEWORK 2 SOLUTIONS 3 By hypothesis, u · v = 0 and u · w = 0. Since we also know u · u = kuk2 and u · 0 = 0, our equation above gives c1 kuk2 = 0. By hypothesis, u is a nonzero vector, so kuk2 6= 0. The above equation therefore implies c1 = 0. If we repeat this process by dotting the original equation with v, we will similarly find c2 = 0. Dotting the original equation with w gives c3 =0. Thus, c1 = c2 = c3 = 0, and hence the set of vectors {u, v, w} is linearly independent. 4.11 We simply expand the product and use the given data: (u − v + 2w) · (x + y − z) = u · x + u · y − u · z −v·x−v·y+v·z + 2w · x + 2w · y − 2w · z = 1 + 2 − 0 − (−1) − 2 + 3 + 2(1) + 2(−2) − 2(−1) = 5. 7 2 · 3 + 0 · −4 + 1 · 1 3 2 0 1 7.1 3 −4 2 −4 = 3 · 3 + −4 · −4 + 2 · 1 = 27 . 8 5 · 3 + 1 · −4 + −3 · 1 1 5 1 −3 7.3 This product is not defined, since the matrix A has three columns but the vector v only has two rows. 7.4 Again this product is not defined, since the matrix A has four columns but the vector v only has two rows. Remember: the matrix-vector product is only defined with the number of columns of the matrix A equals the number of rows of the vector v. 10.16 The first quadrant is not a subspace of R2, since it is not closed under scalar 1 multiplication. For example, notice that is in the first quadrant, but 2 1 −1 − = is not in the first quadrant. 2 −2 10.21 We must check the three properties of a subspace. In each case, we use the fact that V and W are subspaces. (1) Since V and W are subspaces, we know 0 ∈ V and 0 ∈ W . Thus, 0 ∈ V ∩ W. (2) Suppose x, y ∈ V ∩ W . Then x, y ∈ V , and V is a subspace, so x + y ∈ V . Similarly, x, y ∈ W , and W is a subspace, so x + y ∈ W . Thus, x + y ∈ V ∩ W . (3) Suppose x ∈ V ∩ W , and c ∈ R. Then x ∈ V , and V is a subspace, so cx ∈ V . Similarly, x ∈ W and W is a subspace, so cx ∈ W . Thus, cx ∈ V ∩ W . 10.22 Again, we must check the three properties of a subspace. (1) Since V and W are subspaces, we know 0 ∈ V and 0 ∈ W . Therefore, 0 + 0 ∈ V + W , and hence 0 ∈ V + W . 4 HOMEWORK 2 SOLUTIONS (2) Suppose x, y ∈ V +W . Then we can write x = v1 +w1 and y = v2 +w2 for some v1 , v2 ∈ V , w1 , w2 ∈ W . But then x + y = (v1 + v2 ) + (w1 + w2 ). Since V is a subspace, v1 + v2 ∈ V ; since W is a subspace, w1 + w2 ∈ W . So, if we write v3 = v1 + v2 and w3 = w1 + w2 , then x + y = v3 + w3 ∈ V + W . (3) Suppose x ∈ V + W , and c ∈ R. Write x = v + w for some v ∈ V , w ∈ W . Then v ∈ V , and V is a subspace, so cv ∈ V . Similarly, w ∈ W and W is a subspace, so cw ∈ W . Thus, cx = (cv) + (cw) ∈ V + W . 11.14 By hypothesis, {v1 , v2 , v3 } is a basis for V . By our handy theorem (Proposition 12.3), to check that another set of three vectors is also a basis for V , we only have to check either: 1) the set spans V ; or 2) the set is linearly independent. So, to prove that {2v1 − v2 , v1 + v2 + 2v3 , v1 + v3 }, we’ll simply check that the set is linearly independent. Suppose c1 (2v1 − v2 ) + c2 (v1 + v2 + 2v3 ) + c3 (v1 + v3 ) = 0 for some ci ∈ R. We must show that c1 = c2 = c3 = 0. Rewrite the above equation as (2c1 + c2 + c3 )v1 + (−c1 + c2 )v2 + (2c2 + c3 )v3 = 0. Since {v1 , v2 , v3 } is linearly independent, the above must be the trivial combination; i.e. we must have 2c1 + c2 + c3 = 0 −c1 + c2 = 0 2c2 + c3 = 0 Solving the last two equations give c1 = c2 and c3 = −2c2 . Plugging these into the first equation gives c2 = 0, and hence c1 = 0 and c3 = 0. 0 0 1 1 0 , , 0 is linearly inde12.12 (a) Sometimes true. The set 1 0 0 0 0 0 1 0 1 0 1 , , 1 is linearly dependent. pendent, but the set 0 0 0 0 0 0 4 (b) Never true. If a set of three vectors spanned R , then by our handy fact 4 (Proposition 12.1), any set of more than three would vectors in R be 1 0 0 0 1 0 linearly dependent. But the standard basis 0 , 0 , 1 , 0 0 0 is a set of four linearly independent vectors. (c) Never true. Since the standard basis for R3 is a set of three vectors spanning R3 , by our handy fact (Prop. 12.1), any set of more than three vectors in R3 must be linearly dependent. 0 0 0 1 HOMEWORK 2 SOLUTIONS 5 0 1 0 1 (d) Sometimes true. The set 0 , 1 , 0 , 1 spans R3 , 0 0 1 1 1 1 1 1 whereas the set 0 , 0 , 0 , 0 clearly does not. 0 0 0 0 (e) Always true, by Proposition 12.3 (since dim(R4 ) = 4). (f) Always true, again by Proposition 12.3. 13.1 We check the two properties required for a function to be a linear transformation: x1 y1 (1) Suppose , ∈ R2 . Then x2 y2 x1 y1 x1 + y1 f + =f x2 y2 x2 + y2 3(x1 + y1 ) = −2(x1 + y1 ) + 5(x2 + y2 ) 3x1 + 3y1 = (−2x1 + 5x2 ) + (−2y1 + 5y2 ) 3x1 3y1 = + −2x1 + 5x2 −2y1 + 5y2 x1 y1 =f +f . x2 y2 x1 (2) Suppose ∈ R2 and c ∈ R. Then x2 x1 cx1 3cx1 f c =f = x2 cx2 −2cx1 + 5cx2 3x1 =c −2x1 + 5x2 x1 = cf . x2 13.3 This function is not a linear transformation. For example, observe that 2 7 f = , 1 −8 but f 2 4 22 7 2 =f = 6= 2 . 1 2 −16 −8 13.20 Take any vector w ∈ T(V ). Then w = T(v) for some v ∈ V . Since the set {v1 , . . . , vk } spans V , we can write v = c1 v1 + · · · + vk for some ci ∈ R. But then observe that w = T(v) = T(c1 v1 + · · · + vk ) = c1 T(v1 ) + · · · + ck T(vk ), 6 HOMEWORK 2 SOLUTIONS and so w ∈ span(T(v1 ), . . . , T(vk )). Thus, the set {T(v1 ), . . . , T(vk )} spans T(V ). 13.22 Since the set {v1 , . . . , vk } is linearly dependent, there exists some ci ∈ R, not all zero, such that c1 v1 + · · · + vk = 0. Applying the linear transformation T to both sides yields T(c1 v1 + · · · + vk ) = T(0), which we can rewrite as c1 T(v1 ) + · · · + ck T(vk ) = 0. Since the ci are not all zero, this implies the set {T(v1 ), . . . , T(vk )} is linearly dependent.