Math 316 Carter Test 2 Spring 2011 General Instructions: Do not write on this test. Write your name on only the outside of your blue books. Do all your work inside your blue books. On the inside of your blue book, write neat complete answers to the questions below. Show your work. This test has nothing to do with Keanu Reeves. Good luck. 1. (20 points) Consider the ordered orthonormal basis √ √ √ √ B = {Q1 , Q2 } = {[1/ 5, 2/ 5]t , [−2/ 5, 1/ 5]t }. Let x0 , y 0 be the B coordinates of a point in the plane. Write the equation of the ellipse 25(x0 )2 + 4(y 0 )2 = 100 in standard form. 2. (10 points) Let A ∈ M (2, 2) denote a symmetric (2 × 2)-matrix (so A = At ). A consequence of the spectral theorem is that QD = AQ where Q is a (2 × 2) orthogonal matrix and D is a diagonal matrix whose non-zero entries are the (real)-eigenvalues of A. Describe this result in terms of a square, a rectangle, a rotation, and possibly a reflection. For convenience assume that A is non-singular. How can you avoid incorporating a reflection? 3. (5 points) Let A ∈ M (n, n) denote a symmetric matrix. Prove that for all vectors X, Y ∈ M (n, 1), (AX) · Y = X · (AY ). 4. (10 points) Use the result of the previous problem to show that if X and Y are eigenvectors for a symmetric matrix A with distinct eigenvalues, then X · Y = 0. [Hint: Let AX = λX, and consider (λX) · Y ]. " # a b 5. (20 points) Let A = denote a (2 × 2) symmetric matrix. By directly comb d puting roots of the characteristic polynomial (quadratic formula) for A, show that the eigenvalues of A are real. [Hint: The sum of two perfect squares is positive]. 1 6. Let A ∈ M (m, n) denote an (m × n)-matrix with rank(A) = r. (a) (5 points) Prove that B = At A is symmetric. (b) (10 points) Prove that rank(B) = r. (c) (10 points) Prove that the eigenvalues of B are positive real numbers. (You may assume the spectral theorem and that B = U EU t for E ∈ M [r, r] and U a partially orthogonal (n × r)-matrix. 7. (10 points) Give a complete statement of either the reduced singular value theorem or the spectral theorem. 2