Introduction to Linear algebra-Strang-Solutions-Manual_ver13

# 31 a x t ax d ax t x d x t at x d t t problem set

This preview shows page 1. Sign up to view the full content.

This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: 01 3 3 If A D 0 then all  D 0 so all  D 0 as in A D . If A is symmetric then 00 A3 D Qƒ3 QT D 0 requires ƒ D 0. The only symmetric A is Q 0 QT D zero matrix. 4  D 10 and 5 6 7 8 9 If  is complex then  is also an eigenvalue .Ax D x /. Always  C  is real. The 10 11 12 13 trace is real so the third eigenvalue of a 3 by 3 real matrix must be real. If x is not real then  D x T Ax =x T x is not always real. Can’t assume real eigenvectors! &quot; # &quot; #        1 1 1 1 31 9 12 :64 :48 :36 :48 2 2 2 2 D2 C4 1 1 ; D0 C25 1 1 13 12 16 :48 :36 :48 :64 2 2 2 2 &quot;# xT 1 Œ x 1 x 2  is an orthogonal matrix so P1 C P2 D x 1 x T C x 2 x T D Œ x 1 x 2  D I; 1 2 xT 2 T T P1 P2 D x 1 .x1 x2 /x2 D 0. Second proof: P1 P2 D P1 .I P1 / D P1 P1 D 0 since 2 P1 D P1 .       0b A0 0A AD has  D i b and i b . The block matrices and are b0 0A A0 also skew-symmetric with  D i b (twice) and  D i b (twice). Solutions to Exercises 68 14 M is skew-symmetric and orthogonal; ’s must be i , i , 15 A D  i 1  i , i to have trace zero. 1 has  D 0; 0 and only one independent eigenvector x D .i; 1/. The i T good property for complex matrices is not AT D A (symmetric) but A D A (Hermitian with real eigenvalues and orthogonal eigenvectors: see Problem 20 and Section 10:2). 16 (a) If Az D y and AT y D z then BŒ y I z  D Œ AzI AT y  D Œ y I z . So  is also an eigenvalue of B . (b) AT Az D AT .y / D 2 z. (c)  D 1, 1, 1, 1; x 1 D .1; 0; 1; 0/, x 2 D .0; 1; 0; 1/, x 3 D .1; 0; 1; 0/, x 4 D .0; 1; 0; 1/. &quot; # &quot;# 001 1 p p 1, 17 The eigenvalues of B D 0 0 1 are 0; 2; 2 by Problem 16 with x 1 D 13 0 1 0 2 3 2 1 1 1 1 x 2 D 4 p 5, x 3 D 4 p 5. 2 2 18 1. y is in the nullspace of A and x is in the column space D row space because A D AT . Those spaces are perpendicular so y T x D 0. 2. If Ax D x and Ay D ˇ y then shift by ˇ : .A ˇI /x D . ˇ /x and .A ˇI /y D 0 and again x ?y . &quot; # &quot; # Perpendicular for A 1 10 10 1 1 0 ; B has S D 0 1 0 . Not perpendicular for B 19 A has S D 1 0 01 0 0 2d since B T ¤ B   T 1 3 C 4i 20 A D is a Hermitian matrix .A D A/. Its eigenvalues 6 and 4 are 3 4i 1 T real. Adjust equations .1/–.2/ in the text to prove that  is always real when A D A: T Ax D x leads to Ax D x : Transpose to x T A D x T  using A D A: Then x T Ax D x T x and also x T Ax D x T x : So  D  is real:  12 21 (a) False. A D 01  (b) True from AT D QƒQT (c) True from A 1 D Qƒ 1 QT (d) False!  01 has 10 1 D i and 2 D i with x 1 D .1; i / ﬁrst for A but x 1 D .1; i / ﬁrst for AT . 22 A and AT have the same ’s but the order of the x ’s can change. A D  23 A is invertible, orthogonal, permutation, diagonalizable, Markov; B is projection, di- agonalizable, Markov. A allows QR; SƒS 1 ; QƒQT ; B allows SƒS 24 Symmetry gives QƒQT if b D 1; repeated  and no S if b D 1 and QƒQT . 1; singular if b D 0. 25 Orthogonal and symmetric requires jj D    real, so   ˙1. Then A D ˙I  1 and D or  A D QƒQT D cos  sin...
View Full Document

## This note was uploaded on 09/25/2012 for the course PHY 103 taught by Professor Minki during the Spring '12 term at Korea Advanced Institute of Science and Technology.

Ask a homework question - tutors are online