Linear Algebra and Geometry. Written test of september 12, 2013.
1. Let P = (1, 2, 1) and let R be the line {(1, 5, −1) + t(1, 1, −1)}. Compute the distance between P and L and find the point of L which is nearest to P .
Solution. Let u = (1, 2, 1) − (1, 5, −1) = (0, −3, 2). We have that the projection of u along L((1, 1, −1) is pr(u) = −5/3(1, 1, −1). Therefore the nearest point is
H = (1, 5, −1) − 5/3(1, 1, −1) The distance is
k u − pr(u) k=k (0, −3, 2) + 5/3(1, 1, −1) k=k 1/3(5, −4, −1) k= √ 42/3
2. A particle moves along the ellipse 2x 2 + y 2 = 1 with position vector r(t) = (x(t), y(t)).
The motion is such that x 0 (t) = −4y(t) for each t.
(a) Prove that y 0 (t) is proportional to x(t) and find the factor of proportionality;
(b) How much time is required for the particle to go once around the ellipse?
Solution. We have that 2x(t) 2 + y(t) 2 = 1 for each t. Differentiating we get that 4x(t)x 0 (t) + 2y(t)y 0 (t) = 0. Since x 0 (t) = −4y(t) we get −16x(t)y(t) + 2(y(t)y 0 (t) = 2y(t)(−8x(t) + y 0 (t)) ≡ 0. Since y(t) is non-zero except for some isolated points of t we conclude, from the continuity of y 0 (t), that y 0 (t) = 8x(t) for each t.
(b) We have that (x 0 (t), y 0 (t) = (−4y(t), 8x(t)) Differentiating another time we get
(x 00 (t), y 00 (t)) = (−32x(t), −32y(t)). It turns out that both x(t) and y(t) are solutions of the equation f 00 = −32f . Therefore they are of the form a sin 4 √
2t + b cos 4 √
2t. Hence the time to go once around the ellipse is T = 2π/(4 √
2).
3. In V 4 , with the usual dot product, let W = L((1, 1, −1, 0), (1, 0, 0, 1)) and let v = (1, 1, 1, 0).
(a) Find a vector w ∈ W and a vector u ∈ W ⊥ such that v = w + u.
(b) Is it possible to find two other vectors w 0 ∈ W and u 0 ∈ W ⊥ , different from w and u, such that v = w 0 + u 0 ?
Solution. (b) By the Theorem on Orthogonal Decompositions, vectors w and u as required do exist, and they are unique. Therefore the answer to (b) is NO.
(a) By the Theorem on Orthogonal Decompositions, we have that w must be the projection of v onto W , which will be denoted pr W (v) , and u = v − pr W (v). Therefore we have only to compute the projection of v onto W . In order to do that we have first to orthogonalize the given basis of W . We take w 1 = (1, 1, −1, 0). Moreover we have that
(1, 0, 0, 1) − pr L(v
1) ((1, 0, 0, 1)) = (1, 0, 0, 1) − (1/3)(1, 1, −1, 0) = (1/3)(2, −1, 1, 3) Therefore, to simplify the calculation, we can take w 2 = (2, −1, 1, 3). Finally, se have that
pr W (v) = pr L(w
1
) (v) + pr L(w
2