Ala2 SSM PDF

Students’ Solutions
Manual
for
Applied Linear Algebra

by
Peter J. Olver
and Chehrzad Shakiban
Second Edition
Undergraduate Texts in Mathematics
Springer, New York, 2018.
ISBN 978–3–319–91040–6
To the Student
These solutions are a resource for students studying the second edition of our text Applied
Linear Algebra, published by Springer in 2018. An expanded solutions manual is available
for registered instructors of courses adopting it as the textbook.
Using the Manual

The material taught in this book requires an active engagement with the exercises, and
we urge you not to read the solutions in advance. Rather, you should use the ones in
this manual as a means of verifying that your solution is correct. (It is our hope that
all solutions appearing here are correct; errors should be reported to the authors.) If you
get stuck on an exercise, try skimming the solution to get a hint for how to proceed, but
then work out the exercise yourself. The more you can do on your own, the more you will
learn. Please note: for students taking a course based on Applied Linear Algebra, copying
solutions from this Manual can place you in violation of academic honesty. In particular,
many solutions here just provide the final answer, and for full credit one must also supply
an explanation of how this is found.
Acknowledgements
We thank a number of people, who are named in the text, for corrections to the solutions
manuals that accompanied the first edition. Of course, as authors, we take full respon-
sibility for all errors that may yet appear. We encourage readers to inform us of any
misprints, errors, and unclear explanations that they may find, and will accordingly up-
date this manual on a timely basis. Corrections will be posted on the text’s dedicated web
site:
http://www.math.umn.edu/∼olver/ala2.html
Peter J. Olver Cheri Shakiban

University of Minnesota University of St. Thomas
olver@umn.edu cshakiban@stthomas.edu
Minnesota, September 2018

Table of Contents
Chapter 1. Linear Algebraic Systems . . . . . . . . . . . . . . . . . 1
Chapter 2. Vector Spaces and Bases . . . . . . . . . . . . . . . . 11
Chapter 3. Inner Products and Norms . . . . . . . . . . . . . . . 19
Chapter 4. Orthogonality . . . . . . . . . . . . . . . . . . . . . 28
Chapter 5. Minimization and Least Squares . . . . . . . . . . . . . 36
Chapter 6. Equilibrium . . . . . . . . . . . . . . . . . . . . . 44
Chapter 7. Linearity . . . . . . . . . . . . . . . . . . . . . . . 49
Chapter 8. Eigenvalues and Singular Values . . . . . . . . . . . . . 58
Chapter 9. Iteration . . . . . . . . . . . . . . . . . . . . . . . 70
Chapter 10. Dynamics . . . . . . . . . . . . . . . . . . . . . . 81

Students’ Solutions Manual for
Chapter 1: Linear Algebraic Systems
1.1.1. (b) Reduce the system to 6 u + v = 5, − 25 v = 5

2 ; then use Back Substitution to solve
for u = 1, v = −1.
(d) Reduce the system to 2 u − v + 2 w = 2, − 32 v + 4 w = 2, − w = 0; then solve for
u = 13 , v = − 43 , w = 0.
(f ) Reduce the system to x + z − 2 w = − 3, − y + 3 w = 1, − 4 z − 16 w = − 4, 6 w = 6; then
solve for x = 2, y = 2, z = −3, w = 1.
♥ 1.1.3. (a) With Forward Substitution, we just start with the top equation and work down.
Thus 2 x = −6 so x = −3. Plugging this into the second equation gives 12 + 3y = 3, and so
y = −3. Plugging the values of x and y in the third equation yields −3 + 4(−3) − z = 7, and
so z = −22.
 
0
 
1.2.1. (a) 3 × 4, (b) 7, (c) 6, (d) ( −2 0 1 2 ), (e)  2
 .
−6
     
1 2 3 1 2 3 4 1
     
1.2.2. Examples: (a) 4 5 6 4 7  2 .
 , (c)  5 6 , (e)  
7 8 9 7 8 9 3 3
! ! !
6 1 u 5
1.2.4. (b) A = , x= , b= ;
3 −2 v 5
     
2 −1 2 u 2
     
(d) A =  −1 −1 3    
, x =  v , b =  1 ;
3 0 −2 w 1
     
1 0 1 −2 x −3
     
 2 −1 2 −1   y  −5 
(f ) A = 

, x =  , b =  .
 0 −6 −4 2

 
 z

 2

1 3 2 −1 w 1
1.2.5. (b) u + w = −1, u + v = −1, v + w = 2. The solution is u = −2, v = 1, w = 1.

(c) 3 x1 − x3 = 1, −2 x1 − x2 = 0, x1 + x2 − 3 x3 = 1.
The solution is x1 = 15 , x2 = − 52 , x3 = − 52 .
   
1 0 0 0 0 0 0 0 0 0
   
0 1 0 0 0 0 0 0 0 0
   

1.2.6. (a) I = 0

0 1 0 0,

O= 
0

0 0 0 0.


0 0 0 1 0

0 0 0 0 0
0 0 0 0 1 0 0 0 0 0
(b) I + O = I , I O = O I = O. No, it does not.
1
c 2018 Peter J. Olver and Chehrzad Shakiban
2 Students’ Solutions Manual: Chapter 1
 
! 1 11 9
3 6 0 
3

1.2.7. (b) undefined, (c) , (f )  −12 −12 
,
−1 4 2
7 8 8
1.2.9. 1, 6, 11, 16.

 
1 0 0
 
1.2.10. (a) 0 0 0
 .
0 0 −1
1.2.11. (a) True.

!
x y
1.2.14. B = where x, y are arbitrary.
0 x
1.2.17. A On×p = Om×p , Ol×m A = Ol×n .

! ! !
1 0 0 0 0 0
1.2.19. False: for example, = .
0 0 1 0 0 0
1.2.22. (a) A must be a square matrix. (b) By associativity, A A2 = A A A = A2 A = A3 .

!
−1 1
1.2.25. (a) X = .
3 −2
♦ 1.2.29. (a) The ith entry of A z is 1 ai1 + 1 ai2 + · · · + 1 ain = ai1 + · · · + ain , which is
1
the ith row sum. (b) Each row of W has n − 1 entries equal to n and one entry equal
1−n 1 1−n
to n and so its row sums are (n − 1) n + n = 0. Therefore, by part (a), W z = 0.
Consequently, the row sums of B = A W are the entries of B z = A W z = A 0 = 0, and the
result follows.
♥ 1.2.34. (a) This follows by direct computation. (b) (i )
! ! ! ! ! ! !
−2 1 1 −2 −2 1 −2 4 1 0 −1 4
= ( 1 −2 ) + (1 0) = + = .
3 2 1 0 3 2 3 −6 2 0 5 −6
11
1.3.1. (b) u = 1, v = −1; (d) x1 = 3 , x2 = − 10 2
3 , x3 = − 3 ;
(f ) a = 13 , b = 0, c = 34 , d = − 23 .
! !
1 7 4 2R1 +R2 1 7 4
1.3.2. (a) −→ . Back Substitution yields x2 = 2, x1 = −10.
−2 −9 2 0 5 10
     
1 −2 1 0 4R +R 1 −2 1 0 3 R +R 1 −2 1 0
  1 3 2 2 3 
(c) 
 0 2 −8 8 
 −→  0
 2 −8 8 
 −→  0
 2 −8 8  .

−4 5 9 −9 0 −3 13 −9 0 0 1 3
Back Substitution yields z = 3, y = 16, x = 29.

Students’ Solutions Manual: Chapter 1 3
   
1 0 −2 0
−1 1 0 −2 0
−1
   


0 1 0 −1

2
0

1 0 −1

2.
(e)  reduces to
 0 −3 2 0

0

0 0 2 −3

6
−4 0 0 7 −5 0 0 0 −5 15
Back Substitution yields x4 = −3, x3 = − 23 , x2 = −1, x1 = −4.
1.3.3. (a) 3 x + 2 y = 2, − 4 x − 3 y = −1; solution: x = 4, y = −5.

! !
2 1 2 1
1.3.4. (a) Regular: −→ 7 .
1 4 0 2
   
1 −2 3 1 −2 3
   
(d) Not regular: 
 −2 4 −1 
 −→ 0
 0 5.
3 −1 2 0 5 −7
! !
−i 1+ i −1 −i 1+ i −1

1.3.5. (a)

−→

;
1− i 1 −3 i 0 1−2i 1−2i
use Back Substitution to obtain the solution y = 1, x = 1 − 2 i .
  
 aij , i > j,  aij , i < j,  aij , i = j,
♦ 1.3.11. (a) Set lij =  , uij =  dij = 
0, i ≤ j, 0, i ≥ j, 0, i 6= j.
     
0 0 0 3 0 0 0 1 −1
     
(b) L = 
 1 0 0, D=
0 −4 0, U =
 0 0 2.
−2 0 0 0 0 5 0 0 0
1.3.15. (a) Add −2 times the second row to the first row of a 2 × n matrix.
(c) Add −5 times the third row to the second row of a 3 × n matrix.
   
1 0 0 0 1 0 0 3
   
0 1 0 0 0 1 0 0
1.3.16. (a)  , (c)  .

0 0 1 0

0 0 1 0
0 0 1 1 0 0 0 1
1.3.20. (a) Upper triangular; (b) both upper and lower unitriangular; (d) lower unitriangular.
   
! ! 1 0 0 −1 1 −1
1 0 1 3 
 −1
  
1.3.21. (a) L = , U= , (c) L =  1 0 
, U =  0 2 0,
−1 1 0 3
1 0 1 0 0 3
   
1 0 0 −1 0 0
   
(e) L = 
 −2 1 0, U =  0
 −3 0.
−1 −1 1 0 0 2
1.3.25. (a) aij = 0 for all i 6= j; (c) aij = 0 for all i > j and aii = 1 for all i.
!
1 1
1.3.27. False. For instance is regular. Only if the zero appear in the (1, 1) position
1 0
does it automatically preclude regularity of the matrix.

   
! 0 −1
−1  
 1 ,
 
 −1 .
1.3.32. (a) x = 2 , (c) x =   (e) x =  
3 0 5
2
! !   !  
5 9
1 0 −1 3 − 11  1 11 
1.3.33. (a) L = , U= ; x1 =  , x2 = , x3 =  .
−3 1 0 11 2 1 3
11 11
   
   

1 0 0 
9 −2 −1  1 −2
       
2
(c) L = 
 − 3 1 0
, U= 0
 − 13 1 ;
3
x1 =  2 ,
  x2 =  −9 .
 
   
2 5 3 −1
9 3 1 0 0 − 13
 5  1 
    4
1 0 0 0 1 0 −1 0    14 
     1  
 0 1 0 0  0 2 3 −1  −  − 5 
 4  14 
(e) L = 
 , U = 
7 ; x1 =  , x2 =  .
 −1
3
2 1 0

0 0 − 27 2
 1
 

 1 

 4  14 
0 − 12 −1 1 0 0 0 4 1 1
4 2
1.4.1. (a) Nonsingular, (c) nonsingular, (e) singular.
1.4.2. (a) Regular and nonsingular, (c) nonsingular.
1.4.5. (b) x1 = 0, x2 = −1, x3 = 2; (d) x = − 13 9

2 , y = − 2 , z = −1, w = −3.
1.4.6. True. All regular matrices are nonsingular.
 
  0 1 0 0
0 1 0  
  1 0 0 0
1.4.11. (a) 
0 0 1, (c)


.
0 0 0 1
1 0 0
0 0 1 0
     
1 0 0 0 1
0 0 0 0
0 0 1
     
10 0 0 0 0 0 1 0 1 0 0
1.4.13. (b)  ,  ,  ,

00 0 1

0 1 0 0

1 0 0 0
0 0 1 0 0 0 1 0 0 0 1 0
     
1 0 0 0 0 0 0 1 0 1 0 0
     
0 1 0 0 1 0 0 0 0 0 0 1
 ,  ,  .

0 0 0 1

0 1 0 0

1 0 0 0
0 0 1 0 0 0 1 0 0 0 1 0
1.4.14. (a) True, since interchanging the same pair of rows twice brings you back to where you
started.
 
0 0 1
 
1.4.16. (a) 0 1 0
 . (b) True.
1 0 0

1.4.19. Let ri , rj denote the rows of the matrix in question. After the first elementary row op-
eration, the rows are ri and rj + ri . After the second, they are ri − (rj + ri ) = − rj and
rj + ri . After the third operation, we are left with − rj and rj + ri + (− rj ) = ri .
! ! ! !  
5
0 1 0 1 1 0 2 −1 2
 ;
1.4.21. (a) = , x=
1 0 2 −1 0 1 0 1 3
       
0 0 1 0 1 −3 1 0 0 1 0 2 −1
       
(c) 1 0 0  2 3  0  1 −3   1
  0  = 0 1  0 , x=  ;
0 1 0 1 0 2 0 2 1 0 0 9 0
       
0 0 1 0 0 1 0 0 1 0 0 0 1 4 −1 2 −1
       
1 0 0 0 2 3 1 0 0 1 0 0 0 1 0 0  −1 
(e)   =  , x =  .

0 1 0 0 
 1 4 −1 2

 2 −5 1 0 
 0

0 3 −4  
 1

0 0 0 1 7 −1 2 3 7 −29 3 1 0 0 0 1 3
     
0 0 1 0 0 1 −1 1 1 0 0 0 1 −1 1 −3
     
0 1 0 0 0 1 1 0 0 1 0 0 0 1 1 0
1.4.22. (b) 

 =  ;
1 0 0 0 
 1 −1 1 −3 


0 1 1 0 
 0 0 −2 1
5 3
0 0 0 1 1 2 −1 1 1 3 2 1 0 0 0 2
solution: x = 4, y = 0, z = 1, w = 1.
1.4.26. False. Changing the permuation matrix typically changes the pivots.
       
2 1 1 3 −1 −1 1 0 0 3 −1 −1 2 1 1
       
1.5.1. (b) 3 2 1  1  0  1  1
   −4 2  = 0 1  =  −4 2 3 2 .
2 1 2 −1 0 1 0 0 1 −1 0 1 2 1 2
 
1 0 0 0
! !  
0 1 1 2 0

1 0 0.
1.5.3. (a) , (c) , (e) 
1 0 0 1 0 −6 1 0
0 0 0 1
!  
2 1
−1 −1 1 3 3
1.5.6. (a) A = , B −1 =  .
2 −1 − 1 1
3 3
!  
1
2 1 0 3
(b) C = , C −1 = B −1 A−1 =  .
3 0 1 − 23
   
0 0 0 1 1 0 0 0
   
0 0 1 0 0 0 1 0
1.5.9. (a) 

, (c)  .
0 1 0 0

0 0 0 1
1 0 0 0 0 1 0 0
!
1 −1 1
1.5.13. Since c is a scalar, A (c A) = c A−1 A = I .
c c

1.5.16. If all the diagonal entries are nonzero, then D−1 D = I . On the other hand, if one of
diagonal entries is zero, then all the entries in that row are zero, and so D is singular.
♦ 1.5.19. (a) A = I −1 A I . (b) If B = S −1 AS, then A = S B S −1 = T −1 B T , where T = S −1 .

 
−1 ! 1 !
1 1  0 1 0

0
1.5.21. (a) B A = 1=  .
−1 −1 1 0 1
1 1
(b) A X = I does not have a solution. Indeed, the first column of this matrix equation is
   
1 −1 ! 1
  x  
the linear system 
0 1 = 
 0 , which has no solutions since x − y = 1, y = 0,
y
1 1 0
and x + y = 0 are incompatible.
 
5 1 5

1 3


− 8 8 8 
− 8 8 
1 1

1.5.25. (b) 
3
; (d) no inverse; (f ) 
 − 2 2 − 21 .

8 − 81 
7

8 − 38 8
1
! ! ! !
1 0 01 1 3 1 3
1.5.26. (b) = ; (d) not possible;
3 1 −8 0 0 1 3 1
     
1 0 0 1 0 0 1 0 0 1 0 0 1 0 0
     
(f ) 
3 1 0 0 1
 0   
  0 1 0   0 −1 0   0 1 0 

0 0 1 2 0 1 0 3 1 0 0 1 0 0 8
     
1 0 3 1 0 0 1 2 0 1 2 3
     
0 1 00 1 4    
    0 1 0  =  3 5 5 .
0 0 1 0 0 1 0 0 1 2 1 2
 
  i 0 −1
i 1
− 2 2  
1.5.28. (a)  , (c) 1− i −i 1 
1  .
2 − 2i −1 −1 −i
    
     9 −15 −8 3 2
5 2
17 17  2  2 

   
1.5.31. (b)  =  , (d)  6 −10 −5    
 −1  =  3 ,
1 3
− 17 17 12 2 −1 2 1 5 0
    
1 0 1 1 4 3
    
0 0 −1 −1   11   1
(f )   = .

2 −1 −1 0  −7   4 
  

2 −1 −1 −1 6 −2
 
1
  8
1  
 
1.5.32. (b)  4 ; (d) singular matrix; (f )  − 1 .
1  2 
 
4 5
8

! ! ! ! !
0 1 0 4 1 0 −7 0 1 − 72
1.5.33. (b) = ,
1 0 −7 2 0 1 0 4 0 1
      
1 0 0 1 1 5 1 0 0 1 0 0 1 1 5
      7
(d) 0 0 1 1 1 −2  = 2 1 0 0 −3 0 0 1 
      3 ,
0 1 0 2 −1 3 1 0 1 0 0 −7 0 0 1
     
1 −1 1 2 1 0 0 0 1 0 0 0 1 −1 1 2
     
1 −4 1 5 1
  1 0 0  0
 −3 0 0  0
 1 0 −1 
(f ) 
 = .
1 2 −1 −1 
 1 −1 1 0 
 0 0 −2 0 
 0 0 1 0
3 1 1 6 3 − 43 1 1 0 0 0 4 0 0 0 1
 7 
 
! 1 3
 
−8    2 
1.5.34. (b) , (d)  
 −2 , (f ) 

.
3  5  
0
− 35
 
! 1 2 !
1 0 

 1 3 5
1.6.1. (b) , (d)  2 0, (f ) .
1 2 2 4 6
−1 2
 
3 1 !
  −1 2 −3
1.6.2. AT = 
 −1 2, BT = ,
2 0 4
−1 1
 
! −1 6 −5
−2 0  
(A B)T = B T AT = , (B A)T = AT B T = 
 5 −2 11 
.
2 6
3 −2 7
1.6.5. (A B C)T = C T B T AT
1.6.8. (a) (A B)−T = ((A B)T )−1 = (B T AT )−1 = (AT )−1 (B T )−1 = A−T B −T .
! !
1 0 1 −2
(b) A B = , so (A B)−T = ,
2 1 0 1
! ! !
−T 0 −1 −T 1 −1 −T −T 1 −2
while A = ,B = , so A B = .
1 1 −1 2 0 1
♦ 1.6.13. (a) Using Exercise 1.6.12, aij = eT T

i A ej = ei B ej = bij for all i, j.
! ! ! !
1 2 1 1 0 0 0 −1
(b) Two examples: A = , B= ; A= , B= .
0 1 1 1 0 0 1 0
1.6.17. (b) a = −1, b = 2, c = 3.
♦ 1.6.20. True. Invert both sides of the equation AT = A, and use Lemma 1.32.
! ! ! !
1 1 1 0 1 0 1 1
1.6.25. (a) = ,
1 4 1 1 0 3 0 1
     
1 −1 −1 1 0 0 1 0 0 1 −1 −1
     1
(c)  −1 3 2 =  −1 1 0 0 2 0 0 1 
     2 .
1
−1 2 0 −1 2 1 0 0 − 32 0 0 1

♦ 1.6.30. Write A = L D V , then AT = V T D U T = L eU e = V T and U

e , where L e Thus, AT
e = D L.
e , which are the pivots of AT , are the same as those
is regular since the diagonal entries of U
of D and U , which are the pivots of A.
1.7.1. (b) The solution is x = −4, y = −5, z = −1. Gaussian Elimination and Back Substitu-
tion requires 17 multiplications and 11 additions; Gauss–Jordan uses 20 multiplications and
 
0 −1 −1
 
11 additions; computing A−1 =  
 2 −8 −5  takes 27 multiplications and 12 additions,
3
2 −5 −3
while multiplying A−1 b = x takes another 9 multiplications and 6 additions.
♦ 1.7.4. We begin by proving (1.63). We must show that 1 + 2 + 3 + . . . + (n − 1) = 21 n(n − 1)

for n = 2, 3, . . .. For n = 2 both sides equal 1. Assume that (1.63) is true for n = k. Then
1 + 2 + 3 + . . . + (k − 1) + k = 21 k(k − 1) + k = 21 k(k + 1), so (1.63) is true for n = k + 1. Now
1
the first equation in (1.64) follows if we note that 1 + 2 + 3 + . . . + (n − 1) + n = 2 n(n + 1).
      
1 2 0 1 0 0 1 2 0 −2
      
1.7.9. (a)  −1 −1  
1  =  −1 1 
00 1 1  3
 , x=  .
0 −2 3 0 −2 1 0 0 5 0
1.7.11. For n = 4: (a)

    
2 −1 0 0 1 0 0 0 2 −1 0 0
   1  3 
 −1 2 −1 0 − 1 0 0 0 −1 0
 = 2  2  ,

 0 −1 2 −1 


 0 − 32 1 0 
 0 0 4
3 −1 

0 0 −1 2 0 0 − 43 1 0 0 0 5
4
T
(b) ( 2, 3, 3, 2 ) . (c) xi = i(n − i + 1)/2 for i = 1, . . . , n.
    

4 1 1 
1 0 0  4 1 1
    
♠ 1.7.13. 1 1 1 0 15 3
 4 = 4 1  0 4 4
    
1 1 18
1 1 4 4 5 1 0 0 5
! ! !
−8 −10 −8.1
1.7.16. (a) , (b) , (c) . (d) Partial pivoting reduces the effect of
4 −4.1 −4.1
round off errors and results in a significantly more accurate answer.
   
6   0
 5  1.2  
    1
1.7.20. (a)  − 13  =  −2.6 
  .
 5  , (c)  
  1
−1.8
− 95 0

 2 
 
   121  .0165
1 !  38   
− 13 −.0769    .3141 
   121   .
1.7.21. (a) = , (c)   = 
8 .6154 
 59 
  .2438 

13  242 
56 −.4628
− 121
1.8.1. (a) Unique solution: (− 21 , − 34 )T ; (c) no solutions;

(e) infinitely many solutions: (5 − 2 z, 1, z, 0)T , where z is arbitrary.
1.8.2. (b) Incompatible; (c) (1, 0)T ;

(d) (1 + 3 x2 − 2 x3 , x2 , x3 )T , where x2 and x3 are arbitrary.
T
1.8.5. (a) 1 + i − 21 (1 + i )y, y, − i , where y is arbitrary.
1.8.7. (b) 1, (d) 3, (f ) 1.

! ! !
2 1 3 1 0 2 1 3
1.8.8. (b) = ,
−2 −1 −3 −1 1 0 0 0
     
1 0 0 2 −1 0 1 0 0 2 −1 0
   1  3 
(d) 0 0 11
 1 −1  
=2 1 0 0
 2 −1 
,
0 1 0 2 −1 1 1 0 1 0 0 1
(f ) ( 0 −1 2 5 ) = ( 1 )( 0 −1 2 5 ).
1.8.11. Example: (a) x2 + y 2 = 1, x2 − y 2 = 2.
1.8.13. True. For example, take a matrix in row echelon form with r pivots, e.g., the matrix A
with aii = 1 for i = 1, . . . , r, and all other entries equal to 0.
1.8.17. 1.
♦ 1.8.21. By Proposition 1.39, A can be reduced to row echelon form U by a sequence of elemen-
tary row operations. Therefore, as in the proof of the L U decomposition, A = E1 E2 · · · EN U
where E1−1 , . . . , EN
−1
are the elementary matrices representing the row operations. If A is
singular, then U = Z must have at least one all-zero row.
1.8.22. (a) x = z, y = z, where z is arbitrary; (c) x = y = z = 0.

T
1
1.8.23. (a) 3 y, y , where y is arbitrary;
T
(c) − 11 3 2 6
5 z + 5 w, 5 z − 5 w, z, w , where z and w are arbitrary.
1.8.27. (b) k = 0 or k = 21 .

!
2 −1
1.9.1. (a) Regular matrix, reduces to upper triangular form U = , so its determinant is 2.
0 1
 
−1 0 3
 
(b) Singular matrix, row echelon form U =  0 1 −2 
 , so its determinant is 0.
0 0 0
 
−2 1 3
 
(d) Nonsingular matrix, reduces to upper triangular form U =  0 1 −1 
  after one row
interchange, so its determinant is 6. 0 0 3
1.9.4. (a) True. By Theorem 1.52, A is nonsingular, so, by Theorem 1.18, A−1 exists.
! !
2 3 0 1
(c) False. For A = and B = , we have
−1 −2 0 0
!
2 4
det(A + B) = det = 0 6= −1 = det A + det B.
−1 −2
1.9.5. By (1.85, 86) and commutativity of numeric multiplication,

1
det B = det(S −1 A S) = det S −1 det A det S = det A det S = det A.
det S
♦ 1.9.9.
! !
a b a b
det = a d + a k b − b c − b k a = a d − b c = det ,
c + ka d + kb c d
! !
c d a b
det = c b − a d = −(a d − b c) = − det ,
a b c d
! !
ka kb a b
det = k a d − k b c = k (a d − b c) = k det ,
c d c d
!
a b
det = a d − b 0 = ad.
0 d
1.9.13.
 
a11 a12 a13 a14
 
 a21 a22 a23 a24 
det 

=
 a31 a32 a33 a34 

a41 a42 a43 a44
a11 a22 a33 a44 − a11 a22 a34 a43 − a11 a23 a32 a44 + a11 a23 a34 a42 − a11 a24 a33 a42
+ a11 a24 a32 a43 − a12 a21 a33 a44 + a12 a21 a34 a43 + a12 a23 a31 a44 − a12 a23 a34 a41
+ a12 a24 a33 a41 − a12 a24 a31 a43 + a13 a21 a32 a44 − a13 a21 a34 a42 − a13 a22 a31 a44
+ a13 a22 a34 a41 − a13 a24 a32 a41 + a13 a24 a31 a42 − a14 a21 a32 a43 + a14 a21 a33 a42
+ a14 a22 a31 a43 − a14 a22 a33 a41 + a14 a23 a32 a41 − a14 a23 a31 a42 .
♥ 1.9.20. (a) By direct substitution:

pd − bq aq − pc pd − bq aq − pc
ax + by = a +b = p, cx + dy = c +d = q.
ad − bc ad − bc ad − bc ad − bc
! !
1 13 3 1 1 13
(b) (i ) x = − det = −2.6, y=− det = 5.2.
10 0 2 10 4 0

Chapter 2: Vector Spaces and Bases
2.1.1. Commutativity of Addition:

(x + i y) + (u + i v) = (x + u) + i (y + v) = (u + i v) + (x + i y).
Associativity of Addition:
h i h i
(x + i y) + (u + i v) + (p + i q) = (x + i y) + (u + p) + i (v + q)
= (x + u + p) + i (y + v + q)
h i h i
= (x + u) + i (y + v) + (p + i q) = (x + i y) + (u + i v) + (p + i q).
Additive Identity: 0 = 0 = 0 + i 0 and

(x + i y) + 0 = x + i y = 0 + (x + i y).
Additive Inverse: − (x + i y) = (− x) + i (− y) and
h i h i
(x + i y) + (− x) + i (− y) = 0 = (− x) + i (− y) + (x + i y).
Distributivity:
(c + d) (x + i y) = (c + d) x + i (c + d) y = (c x + d x) + i (c y + d y)
= c (x + i y) + d (x + i y),
c[ (x + i y) + (u + i v) ] = c (x + u) + (y + v) = (c x + c u) + i (c y + c v)
= c (x + i y) + c (u + i v).
Associativity of Scalar Multiplication:
c [ d (x + i y) ] = c [ (d x) + i (d y) ] = (c d x) + i (c d y) = (c d) (x + i y).
Unit for Scalar Multiplication: 1 (x + i y) = (1 x) + i (1 y) = x + i y.
Note: Identifying the complex number x + i y with the vector ( x, y )T ∈ R 2 respects the opera-
tions of vector addition and scalar multiplication, and so we are in effect reproving that R 2 is a
vector space.
2.1.4. (a) ( 1, 1, 1, 1 )T , ( 1, −1, 1, −1 )T , ( 1, 1, 1, 1 )T , ( 1, −1, 1, −1 )T . (b) Obviously not.
2.1.6. (a) f (x) = − 4 x + 3.

! ! !
x−y ex 1
2.1.7. (a) , , and , which is a constant function.
xy cos y 3
! !
x − y + ex + 1 − 5 x + 5 y − 5 ex − 5
(b) Their sum is . Multiplied by − 5 is .
x y + cos y + 3 − 5 x y − 5 cos y − 15
!
0
(c) The zero element is the constant function 0 = .
0
2.1.11. (j) Let z = c 0. Then z + z = c (0 + 0) = c 0 = z, and so, as in the proof of (h), z = 0.
11
2.2.2. (a) Not a subspace; (c) subspace; (e) not a subspace; (g) subspace.
-1
-0.5 -1
0 -0.5
0.5 0
1
0.5 -1.5
1 -1.75
-2
-2.25
2 -2.5
1.5
1.25
2.2.3. (a) Subspace: 0

(e) Not a subspace: 1
0.75
-2
0.5
-1
-0.5
0
0.5
1
2.2.7. (b) Not a subspace; (d) subspace.
2.2.10. (a) Not a subspace; (c) not a subspace; (f ) subspace;.
2.2.14. (a) Vector space; (c) vector space;

(e) not a vector space: If f is non-negative, then −1 f = − f is not (unless f ≡ 0).
2.2.15. (b) Subspace; (d) subspace, (e) not a subspace.
2.2.16. (a) Subspace; (c) not a subspace: the zero function does not satisfy the condition;
(e) subspace; (g) subspace.
2.2.20. ∇ · (c v + d w) = c ∇ · v + d ∇ · w = 0 whenever ∇ · v = ∇ · w = 0 and c, d, ∈ R.
♦ 2.2.22. (a) If v, w ∈ W ∩ Z, then v, w ∈ W , so c v + d w ∈ W because W is a subspace, and

v, w ∈ Z, so c v + d w ∈ Z because Z is a subspace, hence c v + d w ∈ W ∩ Z.
♥ 2.2.24. (b) Since the only common solution to x = y and x = 3 y is x = y = 0, the lines only
! ! !
x a 3b
intersect at the origin. Moreover, every v = = + , where a = − 21 x + 23 y,
y a b
1 1
b= 2 x− 2 y, can be written as a sum of vectors on each line.
♦ 2.2.30. (a) By induction, we can show that, for n ≥ 1 and x > 0,

Qn−1 (x) − 1/x2
f (n) (x) = e ,
x2 n
where Qn−1 (x) is a polynomial of degree n − 1. Thus,
Qn−1 (x) − 1/x2
lim f (n) (x) = lim e = Qn−1 (0) lim y 2 n e− y = 0 = lim f (n) (x),
x → 0+ x → 0+ x2 n y→∞ x → 0−
because the exponential e− y goes to zero faster than any power of y goes to ∞.
     
−1 2 5
     
2.3.1.  2    
  = 2 −1  −  −4 .
3 2 1

     
1 1 0
     
2.3.3. (a) Yes, since  −2  =  1  − 3 1 ;
     
−3 1 0
   
   
3 1 0 2
       
 0 2  −1   0
(c) No, since the vector equation 

=c  +c 
 1   2 
+c 
 3 
 does not have a

 −1  0  3  1
solution. −2 1 0 −1
2.3.4. (a) No, (c) yes, (e) yes.

-1
1
0.5 -0.5
0 0
-0.5 0.5
-1 1
− 35 6 1
2.3.5. (b) The plane z = x− 5 y: 0
-1
2.3.8. (a) They span P (2) since ax2 +bx+c = 21 (a−2b+c)(x2 +1)+ 21 (a−c)(x2 −1)+b(x2 +x+1).
2.3.9. (a) Yes, (c) no, (e) no.

1 4
2.3.10. (a) sin 3 x = cos 3 x − 2 π , (c) 3 cos 2 x + 4 sin 2 x = 5 cos 2 x − tan−1 3 .
√ √
2.3.13. (a) e2 x ; (c) e3 x , 1; (e) e− x/2 cos 2
3
x, e− x/2 sin 2
3
x.
!
2
2.3.15. (a) = 2 f1 (x) + f2 (x) − f3 (x); (b) not in the span;
1
!
1 − 2x
(c) = f1 (x) − f2 (x) − f3 (x).
−1 − x
m
X n
X n
X m
X
♦ 2.3.19. (a) If v = cj vj and vj = aij wi , then v = bi vi where bi = aij cj , or,
j =1 i=1 i=1 j =1
in vector language, b = A c.
(b) Every v ∈ V can be written as a linear combination of v1 , . . . , vn , and hence, by
part (a), a linear combination of w1 , . . . , wm , which shows that w1 , . . . , wm also span V .
2.3.21. (b) Linearly dependent; (d) linearly independent; (f ) linearly dependent.
2.3.24. (a) Linearly dependent; (c) linearly independent; (e) linearly dependent.
2.3.27. Yes, when it is the zero vector.
♦ 2.3.31. (a) Since v1 , . . . , vn are linearly independent,

0 = c1 v1 + · · · + ck vk = c1 v1 + · · · + ck vk + 0 vk+1 + · · · + 0 vn
if and only if c1 = · · · = ck = 0.
! !
1 2
(b) This is false. For example, v1 = , v2 = , are linearly dependent, but the
1 2
subset consisting of just v1 is linearly independent.

2.3.32. (a) They are linearly dependent since (x2 − 3) + 2(2 − x) − (x − 1)2 ≡ 0.
(b) They do not span P (2) .
2.3.33. (b) Linearly independent; (d) linearly independent; (f ) linearly dependent.
♥ 2.3.37. (a) If c1 f1 (x) + · · · + cn fn (x) ≡ 0, then c1 f1 (xi ) + · · · + cn fn (xi ) = 0 at all sample

points, and so c1 f1 + · · · + cn fn = 0. Thus, linear dependence of the functions implies linear
dependence of their sample vectors.
(b) Sampling f1 (x) = 1 and f2 (x) = x2 at x1 = −1 and x2 = 1 produces the linearly
!
1
dependent sample vectors f1 = f2 = .
1
2.4.1. (a) Basis; (b) not a basis (d) not a basis.
2.4.2. (a) Not a basis; (c) not a basis.

    3 1
1 0
    4 4
2.4.5. (a)  0  ,  1 ; (b)  1  ,  0 .
       
0 2 0 1
♥ 2.4.7. (a) (i ) Left-handed basis; (iii ) not a basis.
2.4.8. (b) The condition p(1) = 0 says a + b + c = 0, so
p(x) = (− b − c) x2 + b x + c = b (− x2 + x) + c(− x2 + 1).
Therefore − x2 + x, − x2 + 1 is a basis, and so dim = 2.

     
3 2 0
     
2.4.10. (a)  
 1 , dim = 1; (b)
 0 ,  −1 , dim = 2.
   
−1 1 3
       
1 1 1 1
   √     √ 
   2     2 
1      
   2   0 − 2 
2.4.13. (a) The sample vectors  , 
  
, 
 
, 
 


are linearly independent and
 1   0   −1   0 
   √     √ 
1 − 22 0 2
2
hence form a basis for R 4 — the space of sample functions.

       
0 1 1 1
     √   √ 
1   √  2  √  2
  1  2 + 2  2 − 2 
4 1  2  − 2 
(b) Sampling x produces  1
=  −  −  .
  2
1
 8   0 
 8  0 

2       √ √
3 2 2
4 1 − 2 2
! ! !
1 0 0 1 0 0
2.4.17. (a) E11 = , E12 = , E22 = ; dimension = 3.
0 0 0 0 0 1
♦ 2.4.20. (a) m ≤ n as otherwise v1 , . . . , vm would be linearly dependent. If m = n then v1 , . . . , vn

are linearly independent and hence, by Theorem 2.31 span all of R n . Since every vector in
their span also belongs to V , we must have V = R n .
T
(c) (i ) Example: 1, 1, 1
2 , ( 1, 0, 0 )T , ( 0, 1, 0 )T .

♦ 2.4.21. (a) By Theorem 2.31, we only need prove linear independence. If

0 = c1 A v1 + · · · + cn A vn = A(c1 v1 + · · · + cn vn ), then, since A is nonsingular,
c1 v1 + · · · + cn vn = 0, and hence c1 = · · · = cn = 0.
(b) A ei is the ith column of A, and so a basis consists of the column vectors of the matrix.
! ! ! !
1 0 1 2
♦ 2.4.24. For instance, take v1 = , v2 = , v3 = . Then = 2 v1 + v2 =
0 1 1 1
v1 + v3 . In fact, there are infinitely many different ways of writing this vector as a linear
combination of v1 , v2 , v3 .
! 1!
b1 3 2 .
2.5.1. (a) Image: all b = such that 4 b1 + b2 = 0; kernel spanned by
b2 1
 
  5
b 
− 4 
 1  
(c) Image: all b =    7 .
 b2  such that − 2 b1 + b2 + b3 = 0; kernel spanned by  −8 
 
b3
1
    1   
− 52 1
4 0
  2    
2.5.2. (a)  0 ,  1 
  3 :  0 :
 : plane; (b) 8  line; (e)   point.
1 0 1 0
   
−1 1+t
   
2.5.4. (a) b =  2  2 + t ,
 ; (b) the general solution is x =   where t is arbitrary.
−1 3+t
2.5.7. In each case, the solution is x = x⋆ + z, where x⋆ is the particular solution and z be-
 
  2  5
1 
− 7 
     6
⋆
longs to the kernel: 
(b) x =  −1 
, z= z
 − 17 ;
 (d) x⋆ = 
 1, z = 0.
  2
0 −3
1
T
2.5.8. The kernel has dimension n−1, with basis − rk−1 e1 +ek = − rk−1 , 0, . . . , 0, 1, 0, . . . , 0
for k = 2, . . . n. The image has dimension 1, with basis (1, rn , r2 n . . . , r(n−1)n )T .
     
−2 −1 −6
2.5.12. x⋆1 =  , x⋆2 =  ; x = x⋆1 + 4 x⋆2 =  .
3 1 7
2 2 2
 
1
 
2.5.14. (a) By direct matrix multiplication: A x⋆1 = A x⋆2 = 
 −3 .

5    
1 −4
   
(b) The general solution is x = x⋆1 + t (x⋆2 − x⋆1 ) = (1 − t) x⋆1 + t x⋆2 =  1  + t
   2.
0 −2
2.5.19. False: in general, (A + B) x⋆ = (A + B) x⋆1 + (A + B) x⋆2 = c + d + B x⋆1 + A x⋆2 , and the

third and fourth terms don’t necessarily add up to 0.

♦ 2.5.20. img A = R n , and so A must be a nonsingular matrix.
! ! ! !
1 1 3 −2
2.5.21. (a) image: ; coimage:
; kernel: ; cokernel: .
2 −3 1 1
       
    1 0 1 −3  
1 1         −3
     1   −1   −3   2   
(c) image:       
 1 ,  0 ; coimage:  , 
; kernel: 
 
, 
 
; cokernel:


 1.
 2   −3   1  0
2 3 1
1 2 0 1
2.5.23. (i ) rank = 1; dim img A = dim coimg A = 1, dim ker A = dim coker A = 1;
! !
−2 2
kernel basis: ; cokernel basis: ; compatibility conditions: 2 b1 + b2 = 0;
1 1
! ! !
1 1 −2
example: b = , with solution x = +z .
−2 0 1
(iii ) rank = 2; dim img A = dim coimg A = 2, dim ker A = 0, dim coker A = 1;
 
20

− 13 
  20 3
kernel: {0}; cokernel basis:  3 ; compatibility conditions: − 13 b1 +

 13 
 13 b2 + b3 = 0;
  1  
1 1
   
example: b =  −2 ,
with solution x =  0 .
   
2 0
(v ) rank = 2; dim img A = dim coimg A = 2, dim ker A = 1, dim coker A = 2; kernel
 9  1 
 
−4  4
 
−1 
 1  
  1


   4  −4 
basis:  −1 ;
  cokernel basis: 

, 
 
;

compatibility: − 49 b1 + 1
4 b2 + b3 = 0,
1 

1 
 
 0 

1 0
 
2    
  1 −1
1  6     
4 b1 − 14 b2 + b4 = 0; example: b =      
 , with solution x =  0  + z  −1 .
3
 
0 1
1
     
  1 0 1
1      
   0  1  −3 

2.5.24. (b) dim = 1; basis:  1 ; (d) dim = 3; basis: 

, 
 
,  .
 −3   2   −8 
 

−1
   
2 −3 7
1 0
   
 1   −1 
2.5.26. (b)   
 , 
.

1  0
0 1
2.5.29. Both sets are linearly independent and hence span a three-dimensional subspace of R 4 .
Moreover, w1 = v1 + v3 , w2 = v1 + v2 + 2 v3 , w3 = v1 + v2 + v3 all lie in the span of
v1 , v2 , v3 and hence, by Theorem 2.31(d) also form a basis for the subspace.
2.5.35. We know img A ⊂ R m is a subspace of dimension r = rank A. In particular,

img A = R m if and only if it has dimension m = rank A.

♦ 2.5.38. If v ∈ ker A then A v = 0 and so B A v = B 0 = 0, so v ∈ ker(B A). The first statement

follows from setting B = A.
2.5.41. True. If ker A = ker B ⊂ R n , then both matrices have n columns, and so n − rank A =
dim ker A = dim ker B = n − rank B.
♦ 2.5.44. Since we know dim img A = r, it suffices to prove that w1 , . . . , wr are linearly indepen-
dent. Given
0 = c1 w1 + · · · + cr wr = c1 A v1 + · · · + cr A vr = A(c1 v1 + · · · + cr vr ),
we deduce that c1 v1 + · · · + cr vr ∈ ker A, and hence can be written as a linear combination
of the kernel basis vectors:
c1 v1 + · · · + cr vr = cr+1 vr+1 + · · · + cn vn .
But v1 , . . . , vn are linearly independent, and so c1 = · · · = cr = cr+1 = · · · = cn = 0, which
proves linear independence of w1 , . . . , wr .
2.6.2. (a) (c)
(e) or, equivalently,
 
−1 0 1 0 0
   
−1 0 1 0  −1
 1 0 0 0
   


0 −1 1 0;
 0 −1 0 1 0
2.6.3. (a)  (c)  .
 0 1 0 −1 


 0

−1 0 0 1

0 0 1 −1  0 0 1 −1 0
 
0 0 0 1 −1
   
−1 0
     
0 
 1 
 0

     
 −1   1  −1 
2.6.4. (a) 1 circuit:  ; (c) 2 circuits:  ,  .
     
 −1   0  1
   
1  1  0
 
 
0 1

 
1 −1 0 0
 
1 0 −1 0
 
 
1 0 0 −1 
♥ 2.6.7. (a) Tetrahedron:  

0

1 −1 0

0 1 0 −1 
 
0 0 1 −1
number of circuits = dim coker A = 3, number of faces = 4.
 
−1 1 0 0 0 0
   
−1 1 0 0 
 0 1 −1 0 0 0
  
♥ 2.6.9. (a) (i ) 
 0 1 −1 0, (iii ) 

0 0 1 −1 0 0.


0 1 0 −1  0 1 0 0 −1 0
0 1 0 0 0 −1
(b)
     
−1 1 0 0 0 −1 1 0 0 0 −1 1 0 0 0
     
 0 −1 1 0 0  0 −1 1 0 0  0 1 −1 0 0
 ,  ,  .

 0 0 −1 1 0

 0 0 −1 1 0

 0 1 0 −1 0
0 0 0 −1 1 0 1 0 0 −1 0 1 0 0 −1
♥ 2.6.10. G3 G4
(a)
 
1 −1 0 0
 
  1 0 −1 0
1 −1 0  
 
  1 0 0 −1 
(b) 1 0 −1   .
, 

0

1 −1 0

0 1 −1 0
 1 0 −1 

0 0 1 −1
♥ 2.6.14. (a) Note that P permutes the rows of A, and corresponds to a relabeling of the ver-
tices of the digraph, while Q permutes its columns, and so corresponds to a relabeling of
the edges. (b) (i ) Equivalent, (iii ) inequivalent, (v ) equivalent.

Chapter 3: Inner Products and Norms
3.1.2. (b) No — not positive definite; (d) no — not bilinear; (f ) yes.
3.1.4. (b) Bilinearity:

h c u + d v , w i = 4 (c u1 + d v1 ) w1 + 2 (c u1 + d v1 ) w2 + 2 (c u2 + d v2 ) w1 +
+ 4 (c u2 + d v2 ) w2 + (c u3 + d v3 ) w3
= c (4 u1 w1 + 2 u1 w2 + 2 u2 w1 + 4 u2 w2 + u3 w3 ) +
+ d (4 v1 w1 + 2 v1 w2 + 2 v2 w1 + 4 v2 w2 + v3 w3 )
= c h u , w i + d h v , w i,
h u , c v + d w i = 4 u1 (c v1 + d w1 ) + 2 u1 (c v2 + d w2 ) + 2 u2 (c v1 + d w1 ) +
+ 4 u2 (c v2 + d w2 ) + u3 (c v3 + d w3 )
= c (4 u1 v1 + 2 u1 v2 + 2 u2 v1 + 4 u2 v2 + u3 v3 ) +
+ d (4 u1 w1 + 2 u1 w2 + 2 u2 w1 + 4 u2 w2 + u3 w3 )
= c h u , v i + d h u , w i.
Symmetry:
h v , w i = 4 v1 w1 + 2 v1 w2 + 2 v2 w1 + 4 v2 w2 + v3 w3
= 4 w1 v1 + 2 w1 v2 + 2 w2 v1 + 4 w2 v2 + w3 v3 = h w , v i.
Positivity:
h v , v i = 4 v12 + 4 v1 v2 + 4 v22 + v32 = (2 v1 + v2 )2 + 3 v22 + v32 > 0
for all v = ( v1 , v2 , v3 )T 6= 0,
because it is a sum of non-negative terms, at least one of which is strictly positive.
0.5
3.1.5. (b) Ellipse with semi-axes √1 , √1 :

2 5 -1 -0.5 0.5 1
-0.5
-1
q q
♦ 3.1.7. kcvk = h cv , cv i = c2 h v , v i = | c | k v k.
♦ 3.1.10. (a) Choosing v = x, we have 0 = h x , x i = k x k2 , and hence x = 0.

(b) Rewrite the condition as 0 = h x , v i − h y , v i = h x − y , v i for all v ∈ V . Now use
part (a) to conclude that x − y = 0 and so x = y.
19
♦ 3.1.12. (a) k u + v k2 − k u − v k2 = h u + v , u + v i − h u − v , u − v i

= h u , u i + 2 h u , v i + h v , v i − h u , u i − 2 h u , v i + h v , v i = 4 h u , v i.
3.1.15. Using (3.2), v · (A w) = vT A w = (AT v)T w = (AT v) · w.
3.1.21. (a) h 1 , x i = 21 , k 1 k = 1, k x k = √1 ;
3
r
(c) h x , ex i = 1, k x k = √1 , k ex k = 1 2
2 (e − 1) .
3
r r
2 56
3.1.22. (b) h f , g i = 0, k f k = 3 , kg k = 15 .
Z 1
3.1.23. (a) Yes; (b) no, since it fails positivity: for instance, (1 − x)2 x dx = − 34 .
−1
√ √
3.1.26. No. For example, on [ − 1, 1 ], k 1 k = 2 , but k 1 k2 = 2 6= k 12 k = 2.
♦ 3.1.30. (a) To prove the first bilinearity condition:

Z bh i
h cf + dg , h i = c f (x) + d g(x) h(x) w(x) dx
a
Z b Z b
=c f (x) h(x) w(x) dx + d g(x) h(x) w(x) dx = c h f , h i + d h g , h i.
a a
The second has a similar proof, or follows from symmetry, cf. Exercise 3.1.9.
To prove symmetry:
Z b Z b
hf ,gi = f (x) g(x) w(x) dx = g(x) f (x) w(x) dx = h g , f i.
a a
Z b
As for positivity, h f , f i = f (x)2 w(x) dx ≥ 0. Moreover, since w(x) > 0 and the inte-
a
grand is continuous, Exercise 3.1.29 implies that h f , f i = 0 if and only if f (x)2 w(x) ≡ 0
for all x, and so f (x) ≡ 0.
(b) If w(x0 ) < 0, then, by continuity, w(x) < 0 for x0 − δ ≤ x ≤ x0 + δ for some δ > 0. Now
choose f (x) 6≡ 0 so that f (x) = 0 whenever | x − x0 | > δ. Then
Z b Z x +δ
0
hf ,f i = f (x)2 w(x) dx = f (x)2 w(x) dx < 0, violating positivity.
a x0 −δ
q
3.1.32. (a) h f , g i = 23 , k f k = 1, k g k = 28
45 .
√ √
3.2.1. (a) | v1 · v2 | = 3 ≤ 5 = 5 5 = k v1 k k v2 k; angle: cos−1 53 ≈ .9273;
√ √ √
(c) | v1 · v2 | = 0 ≤ 2 6 = 2 12 = k v1 k k v2 k; angle: 21 π ≈ 1.5708.
√ √
3.2.4. (a) | v · w | = 5 ≤ 7.0711 = 5 10 = k v k k w k.
√ √
3.2.5. (b) | h v , w i | = 11 ≤ 11.7473 = 23 6 = k v k k w k.

♦ 3.2.6. Expanding k v + w k2 − k v − w k2 = 4 h v , w i = 4 k v k k w k cos θ.
kv + wk
kwk
θ
kv − wk
kvk
3.2.9. Set v = ( a1 , . . . , an )T , w = ( 1, 1, . . . , 1 )T , so that Cauchy–Schwarz gives

√ q
| v · w | = | a1 + a2 + · · · + an | ≤ n a21 + a22 + · · · + a2n = k v k k w k.
Equality holds if and only if v = a w, i.e., a1 = a2 = · · · = an .

q q
1 1 2 1
3.2.12. (a) | h f , g i | = 1 ≤ 1.03191 = 3 2 e − 2 = k f k k g k;
q q
(b) | h f , g i | = 2/e = .7358 ≤ 1.555 = 23 12 (e2 − e−2 ) = k f k k g k.
√
2 2
3.2.13. (a) 21 π, (b) cos−1 √ = .450301.
π
q
2 28
3.2.14. (a) | h f , g i | = 3 ≤ 45 = k f k k g k.
3.2.15. (a) a = − 34 ; (b) no.

T
1 7        
3.2.16. All scalar multiples of 2, −4, 1 . 1 2 1 2
       
 −2   −3   −2   −3 
3.2.18. All vectors in the subspace spanned by  ,  , so v = a + b .
       
 1  0  1  0
3.2.21. (a) All solutions to a + b = 1. 0 1 0 1
♦ 3.2.23. Choose v = w; then 0 = h w , w i = k w k2 , and hence w = 0.

Z 1 Z 1
3.2.26. (a) h p1 , p2 i = x− 1
2 dx = 0, h p1 , p3 i = x2 −x+ 1
6 dx = 0,
0 0
Z 1
h p2 , p3 i = x− 1
2 x2 − x + 1
6 dx = 0.
0

3.2.28. p(x) = a (e − 1) x − 1 + b x2 − (e − 2) x for any a, b ∈ R.
√
3.2.32. (a) k v1 + v2 k = 4 ≤ 2 5 = k v1 k + k v2 k,
√ √ √
(c) k v1 + v2 k = 14 ≤ 2 + 12 = k v1 k + k v2 k.
√ √ √
3.2.33. (a) k v1 + v2 k = 5 ≤ 5 + 10 = k v1 k + k v2 k.

r
2 1 2 1 −2
3.2.34. (b) k f + g k = 3 + 2e + 4 e−1 − 2e ≈ 2.40105
r r
2 1 2
≤ 2.72093 ≈ 3 + 2 (e − e−2 ) = k f k + k g k.
r r
133 28
3.2.35. (a) k f + g k = 45 ≈ 1.71917 ≤ 2.71917 ≈ 1 + 45 = k f k + k g k.
Z s s
1 Z 1 Z 1
x 2 x
3.2.37. (a)
0
f (x) g(x) e dx

≤ f (x) e dx g(x)2 ex dx ,
0 0
s s s
Z 1h i2 Z 1 Z 1
x
f (x) + g(x) e dx ≤ f (x)2 ex dx + g(x)2 ex dx ;
0 0 0
r
√
(b) h f , g i = 1
2 (e2 − 1) = 3.1945 ≤ 3.3063 = e − 1 13 (e3 − 1) = k f k k g k,
r r
1 3 7 √
kf + g k = 2
3e +e +e− 3 = 3.8038 ≤ 3.8331 = e − 1 + 13 (e3 − 1) = k f k + k g k;
√
3 e2 − 1
(c) cos θ = q = .9662, so θ = .2607.
2 (e − 1)(e3 − 1)
3.3.1. k v + w k1 = 2 ≤ 2 = 1 + 1 = k v k1 + k w k1 ,
√
k v + w k2 = 2 ≤ 2 = 1 + 1 = k v k2 + k w k2 ,
√
k v + w k3 = 3 2 ≤ 2 = 1 + 1 = k v k3 + k w k3 ,
k v + w k∞ = 1 ≤ 2 = 1 + 1 = k v k∞ + k w k∞ .
3.3.3. (a) k u − v k1 = 5, k u − w k1 = 6, k v − w k1 = 7, so u, v are closest.

√ √ √
(b) k u − v k2 = 13, k u − w k2 = 12, k v − w k2 = 21, so u, w are closest.
1 2 1 1
3.3.6. (a) k f − g k1 = 2 = .5, k f − h k1 = 1 − π = .36338, k g − h k1 = 2 − π = .18169,
r r
1 3 4
so g, h are closest. (b) k f − g k2 = 3 = .57735, k f − h k2 = 2 − π = .47619,
r
5 2
k g − h k2 = 6 − π = .44352, so g, h are closest.
3 5
3.3.7. (a) k f + g k1 = 4 = .75 ≤ 1.3125 ≈ 1 + 16 = k f k1 + k g k1 ;
r r
31 7
(b) k f + g k2 = 48 ≈ .8036 ≤ 1.3819 ≈ 1 + 48 = k f k2 + k g k2 .
3.3.10. (a) Comes from weighted inner product h v , w i = 2 v1 w1 + 3 v2 w2 .

(c) Clearly positive; k c v k = 2 | c v1 | + | c v2 | = | c | 2 | v1 | + | v2 | = | c | k v k;
k v + w k = 2 | v1 + w1 | + | v2 + w2 | ≤ 2 | v1 | + | v2 | + 2 | w1 | + | w2 | = k v k + k w k.
3.3.11. (a) Yes; (b) no, since, for instance, k ( 1, −1, 0 )T k = 0.
3.3.13. True for an inner product norm, but false in general. For example,
k e1 + e2 k1 = 2 = k e1 k1 + k e2 k1 .
Z b Z bh i
♦ 3.3.17. (a) k f + g k1 = | f (x) + g(x) | dx ≤ | f (x) | + | g(x) | dx
a a
Z b Z b
= | f (x) | dx + | g(x) | dx = k f k1 + k g k1 .
a a

   
1 1
 3   3 
   
3.3.20. (b)  2 ; (d)  2 .
 3   3 
   
−1 −1
3.3.22. (b) 2 vectors, namely u = v/k v k and − u = − v/k v k.

1
1 1
0.75
0.5 0.5 0.5
0.25
3.3.24. (a) -1 -0.75-0.5-0.25 0.25 0.5 0.75 1

(c) -1 -0.5 0.5 1
(e) -1 -0.5 0.5 1
-0.25
-0.5 -0.5 -0.5
-0.75
-1 -1
-1
3.3.25.
(a) Unit octahedron:
(c) Ellipsoid with semi-axes √1 , 1, √1 :

2 3
18
3.3.28. (a) 5 x − 65 ; (c) 3
2 x − 21 ; (e) 3
√ x− 1
√ .
2 2 2 2
3.3.29. (a) Yes, (c) yes, (e) no, (g) no — its norm is not defined.
√ √ √
3.3.31. (a) k v k2 = 2 , k v k∞ = 1, and √1 2 ≤ 1 ≤ 2;
2
1
(c) k v k2 = 2, k v k∞ = 1, and 2 2 ≤ 1 ≤ 2.
3.3.32. (a) v = ( a, 0 )T or ( 0, a )T ; (c) v = ( a, 0 )T or ( 0, a )T .

 2
n
X n
X X n
X
3.3.35. (i ) k v k21 =  | vi | = | vi |2 + 2 | vi | | vj | ≥ | vi |2 = k v k22 .
i=1 i=1 i<j i=1
On the other hand, since 2 x y ≤ x2 + y , 2
n
X X n
X
k v k21 = | vi |2 + 2 | vi | | vj | ≤ n | vi |2 = n k v k22 .
i=1 i<j i=1
√ √ √ √
(ii ) (a) k v k2 = 2, k v k1 = 2, and 2 ≤ 2 ≤ 2 2.
(iii ) (a) v = c ej for some j = 1, . . . , n.

3.3.37. In each case, we minimize and maximize k ( cos θ, sin θ )T k for 0 ≤ θ ≤ 2 π:

√ √
(a) c⋆ = 2 , C ⋆ = 3 .
♥ 3.3.40. (a) The maximum (absolute) value of fn (x) is 1 = k fn k∞ . On the other hand,
sZ sZ
∞ n √
k fn k2 = | fn (x) |2 dx = dx = 2n −→ ∞.
−∞ −n
(b) Suppose there exists a constant C such that k f k2 ≤ C k f k∞ for all functions. Then, in
√
particular, 2 n = k fn k2 ≤ C k fn k∞ = C for all n, which is impossible.
3
3.3.45. (a) 4, (c) .9.
! ! !
0 1 1 2 0 −2
3.3.47. False: For instance, if A = ,S= , then B = S −1 A S = , and
0 1 0 1 0 1
k B k∞ = 2 6= 1 = k A k∞ .
♦ 3.3.48. (i ) The 1 matrix norm is the maximum absolute column sum:

 
n
 X 

k A k1 = max  | aij | 1 ≤ j ≤ n .

i=1
5 8
(ii ) (a) 6, (c) 7.
3.4.1. (a) Positive definite: h v , w i = v1 w1 + 2 v2 w2 ; (c) not positive definite;

(e) positive definite: h v , w i = v1 w1 − v1 w2 − v2 w1 + 3 v2 w2 .
!
1 2
♦ 3.4.5. (a) kii = eT
i K ei > 0. (b) For example, K = is not positive definite or even
! 2 1
1 0
semi-definite. (c) For example, K = .
0 0
! ! !
3 1 1 1 4 7
3.4.8. For example, = is not even symmetric. Even the associated
1 1 1 4 2 5
! !
4 7 x
quadratic form ( x y ) = 4 x2 + 9 x y + 5 y 2 is not positive definite.
2 5 y
♦ 3.4.12. If q(x) = xT K x with K T = K, then

q(x + y) − q(x) − q(y) = (x + y)T K(x + y) − xT K x − yT K y = 2 xT K y = 2 h x , y i.
! ! ! !
−1 1 −2 1 −1
3.4.17. xT K x = ( 1 1 ) = 0, but K x = = 6= 0.
1 −2 3 1 1
1
0.5
♦ 3.4.20. (b) x2 + 4 x y + 3 y 2 = (x + y)(x + 3 y) = 0: -1 -0.5 0.5 1
-0.5
-1

 
! ! 9 6 3
10 6 6 −8 
6

3.4.22. (i ) ; positive definite. (iii ) ; positive definite. (v )  6 0;
6 4 −8 13
3 0 3
 
−1
 
positive semi-definite; null directions: all nonzero scalar multiples of  
 1 .
1
 
! 21 12 9
9 −12 
 12

3.4.23. (iii ) , positive definite; (v )  9 3, positive semi-definite.
−12 21
9 3 6
Note: Positive definiteness doesn’t change, since it only depends upon the linear indepen-
dence of the vectors.
 

1 e−1 1
2 (e2 − 1) 
 
3.4.25. K = 
 e−1 1
2 (e2 − 1) 1
3 (e3 − 1) 

 
1
2 (e2 − 1) 1
3 (e3 − 1) 1
4 (e4 − 1)
is positive definite since 1, ex , e2 x are linearly independent functions.
♦ 3.4.30. (a) is a special case of (b) since positive definite matrices are symmetric.
(b) By Theorem 3.34 if S is any symmetric matrix, then S T S = S 2 is always positive semi-
definite, and positive definite if and only if ker S = {0}, i.e., S is nonsingular. In particular,
if S = K > 0, then ker K = {0} and so K 2 > 0.
♦ 3.4.33. Let h x , y i = xT K y be the corresponding inner product. Then kij = h ei , ej i, and

hence K is the Gram matrix associated with the standard basis vectors e1 , . . . , en .
3.5.1. (a) Positive definite; (c) not positive definite; (e) positive definite.
! ! ! !
1 2 1 0 1 0 1 2
3.5.2. (a) = ; not positive definite.
2 3 2 1 0 −1 0 1
     
3 −1 3 1 0 0 3 0 0 1 − 13 1
   1  14  
(c)  −1
 5 1 
 =  −3 1 0 
 0 3 0 
 0 1 73 ; positive definite.
3 8
3 1 5 1 7 1 0 0 7 0 0 1
 
1

1 2 − 12 
 
3.5.4. K =  1
 2 2 0
; yes, it is positive definite.
 
1
− 2 0 3
3.5.5. (a) (x + 4 y)2 − 15 y 2 ; not positive definite. (b) (x − 2 y)2 + 3 y 2 ; positive definite.
3.5.6. (a) (x + 2 z)2 + 3 y 2 + z 2 .

  
1 0 2 x
  
3.5.7. (a) ( x y z )T 
0 2 4  y ; not positive definite;
 
2 4 12 z
  
1 1 −2 x
  
(c) ( x y z )T 
 1 2 −3   y ; positive definite.
 
−2 −3 6 z
n
X
3.5.10. (b) tr K = kii > 0 since, according to Exercise 3.4.5, every diagonal entry of K is
i=1
positive.
3.5.13. Write S = (S + c I ) + (− c I ) = K + N, where N = − c I is negative definite for any

c > 0, while K = S + c I is positive definite provided c ≫ 0 is sufficiently large.
! ! !
4 −12 2 0 2 −6
3.5.19. (b) = ,
−12 45 −6 3 0 3
√
  √ 
2 √1 √1 
2 1 1 2 √0 0
    √2 2 
(d) 
1 2 1 
=
√1 √3 0

 0 √3 √1

.
2 2  2 6 
1 1 2 √1 √1 √2 0 0 √2
2 6 3 3
! ! !
4 −2 2 0 2 −1
3.5.20. (a) = √ √ , (c) no factorization.
−2 4 −1 3 0 3
√
3.5.21. (b) z12 + z22 , where z1 = x1 − x2 , z2 = 3 x2 ;
r r
√
(d) z12 + z22 + z32 , where z1 = 3 x1 − √1 x2 − √1 x3 , z2 = 53 x2 − √1 x3 , z 3 = 28
5 x3 .
3 3 15
(
1, k even,
3.6.2. ek π i = cos k π + i sin k π = (−1)k =
−1, k odd.
√ 1 i 1 i
3.6.5. (a) i = eπ i /2 ; (b) i = eπ i /4 = √ + √ and e5 π i /4 = − √ − √ .
2 2 2 2
3.6.7. (a) 1/z moves in a clockwise direction around a circle of radius 1/r.
♦ 3.6.9. Write z = r e i θ so θ = ph z. Then Re e i ϕ z = Re (r e i (ϕ+θ) ) = r cos(ϕ + θ) ≤ r = | z |,

with equality if and only if ϕ + θ is an integer multiple of 2 π.
3.6.13. Set z = x + i y, w = u + i v, then z w = (x + i y) · (u − i v) = (x u + y v) + i (y u − x v)

has real part Re (z w) = x u + y v, which is the dot product between ( x, y )T and ( u, v )T .
3.6.16. (b) cos 3 θ = cos3 θ − 3 cos θ sin2 θ, sin 3 θ = 3 cos θ sin2 θ − sin3 θ.
3.6.18. ez = ex cos y + i ex sin y = r cos θ + i r sin θ implies r = | ez | = ex and θ = ph ez = y.
♦ 3.6.22. xa+ i b = xa e i b log x = xa cos(b log x) + i xa sin(b log x).

d λx d µx
♦ 3.6.24. (a) e = e cos ν x + i eµ x sin ν x = (µ eµ x cos ν x − ν eµ x sin ν x) +
dx dx
+ i (µ eµ x sin ν x + ν eµ x cos ν x) = (µ + i ν) eµ x cos ν x + i eµ x sin ν x = λ eλ x .
1 1
3.6.25. (a) 2 x+ 4 sin 2 x, (c) − 14 cos 2 x, (e) 3
8 x+ 1
4 sin 2 x + 1
32 sin 4 x.
3.6.26. (b) Linearly dependent; (d) linearly dependent; (f ) linearly independent.

! !
z iz
3.6.27. False — it is not closed under scalar multiplication. For instance, i = is
z iz
not in the subspace since i z = − i z.
3.6.29. (b) Dimension = 2; basis: ( i − 1, 0, 1 )T , ( − i , 1, 0 )T .

T
(d) Dimension = 1; basis: − 14
5 −
8
5 i, 13 4
5 − 5 i,1 .
     
! ! 2 0 1 + 25 i
2 −1 + i 
 −1 +
 
, 

;



;
3.6.30. (b) Image: , ; coimage:  i   1+ i  kernel:  3i 
−4 3
1 − 2i 3− 3i 1
cokernel: {0}.
3.6.33. (a) Not a subspace; (b) subspace; (d) not a subspace.
3.6.35. (a) Belongs: sin x = − 12 i e i x + 1

2 i e− i x ; (c) doesn’t belong;
(d) belongs: sin2 12 x= 1
2 − 1
2 e ix
− 1 −ix
2e .
3.6.36. (a) Sesquilinearity:

h c u + d v , w i = (c u1 + d v1 ) w1 + 2 (c u2 + d v2 ) w2
= c (u1 w1 + 2 u2 w2 ) + d (v1 w1 + 2 v2 w2 ) = c h u , w i + d h v , w i,
h u , c v + d w i = u1 ( c v1 + d w1 ) + 2 u2 ( c v2 + d w2 )
= c̄ (u1 v 1 + 2 u2 v 2 ) + d̄ (u1 w1 + 2 u2 w2 ) = c̄ h u , v i + d̄ h u , w i.
Conjugate Symmetry:
h v , w i = v1 w1 + 2 v2 w2 = w1 v 1 + 2 w2 v 2 = h w , v i.
Positive definite: h v , v i = | v1 |2 + 2 | v2 |2 > 0 whenever v = ( v1 , v2 )T 6= 0.
3.6.37. (a) No, (b) no, (d) yes.

2 2 2
♦ 3.6.40. (a) k z + w k = h z + w , z + w i = k z k + h z , w i + h w , z i + k w k
= k z k2 + h z , w i + h z , w i + k w k2 = k z k2 + 2 Re h z , w i + k w k2 .
3.6.46. (e) Infinitely many, namely u = e i θ v/k v k for any 0 ≤ θ < 2 π.
3.6.48. (b) (i ) h x + i , x − i i = − 32 + i , k x + i k = k x − i k = √2 ;
3
√
13 4
(ii ) | h x + i , x − i i | = 3 ≤ 3 = k x + i k k x − i k,
k (x + i ) + (x − i ) k = k 2 x k = √2 ≤ √4 = k x + i k + k x − i k.
3 3

Chapter 4: Orthogonality
4.1.1. (a) Orthogonal basis (c) not a basis; (e) orthogonal basis.
4.1.2. (b) Orthonormal basis.
4.1.3. (a) Basis; (c) not a basis; (e) orthonormal basis.
4.1.5. (a) a = ± 1.
4.1.6. a = 2 b > 0.      
1 0 0
     
4.1.9. False. Consider the basis v1 = v2 =  1 ,
v3 =  1 ,
Under the weighted in-  0 .
     
0 0 1
ner product, h v1 , v2 i = b > 0, since the coefficients of a, b, c appearing in the inner product
must be strictly positive.
♥ 4.1.13. (a) The (i, j) entry of AT K A is viT K vj = h vi , vj i. Thus, AT K A = I if and only if

(
1, i = j,
h vi , vj i = and so the vectors form an orthonormal basis.
0, i 6= j,
(b) According to part (a), orthonormality requires AT K A = I , and so K = A−T A−1 =
(A AT )−1 is the Gram matrix for A−1 , and K > 0 since A−1 is nonsingular. This also
proves the uniqueness of the inner product.
! !
1 2 10 −7
(c) A = , K= , with inner product
1 3 −7 5
h v , w i = vT K w = 10 v1 w1 − 7 v1 w2 − 7 v2 w1 + 5 v2 w2 .
4.1.16. h v1 + v2 , v1 − v2 i = k v1 k2 − k v2 k2 = 0 by assumption. Moreover, since v1 , v2 are

linearly independent, neither v1 − v2 nor v1 + v2 is zero, and hence Theorem 4.5 implies
that they form an orthogonal basis for the two-dimensional vector space V .

3
4.1.18. (b) p(t) = c 1 − 2 t for any c.
4.1.22. (a) We compute h v1 , v2 i = h v1 , v3 i = h v2 , v3 i = 0 and k v1 k = k v2 k = k v3 k = 1.

T
(b) h v , v1 i = 57 , h v , v2 i = 11
13 , and h v , v3 i = − 37 7
65 , so ( 1, 1, 1 ) = 5 v1 +
11
13 v2 − 37
65 v3 .
2
7 2 11 2
(c) 5 + 13 + − 37
65 = 3 = k v k2 .
4.1.23. (a) Because h v1 , v2 i = v1T K v2 = 0.

h v , v1 i h v , v2 i
(b) v = v + v = 73 v1 − 1
v2 .
k v1 k2 1 k v2 k2 2 3
!2 !2 2 2
v1 · v v2 · v
(c) + = √7 + − √5 = 18 = k v k2 .
k v1 k k v2 k 3 15
28
T T
(d) u1 = √1 , √1 , u2 = − √2 , √1 .
3 3 15 15
√ 2 √ 2
(e) v = h v , u1 i u1 + h v , u2 i u2 = √7 u1 − 15
3 u2 ; k v k2 = 18 = √7 + − 15
3 .
3 3
h x , p1 i 1 h x , p2 i h x , p3 i 1
4.1.26. = , = 1, = 0, so x = 2 p1 (x) + p2 (x).
k p1 k2 2 k p2 k2 k p3 k2
4.1.29. (b) cos x sin x = 1

2 sin 2 x, (d) cos2 x sin3 x = 18 sin x + 1 1
16 sin 3 x − 16 sin 5 x.
Z π Z π (
1 1 1, k = l,
♦ 4.1.31. h e i k x , e i l x i = e i k x e i k x dx = e i (k−l) x dx =
2π −π 2π −π 0, k 6= l.
4.2.1. (b) √1 ( 1, 1, 0 )T , √1 ( −1, 1, −2 )T , √1 ( 1, −1, −1 )T .

2 6 3
T T T T
4.2.2. (a) √1 , 0, √1 , 0 , 0, √1 , 0, − √1 , 1 1 1 1
− 12 , 12 , 12 , 1
2 2 2 2 2, 2, −2, 2 , 2 .
T
4.2.4. (b) Starting with the basis 1
2 , 1, 0 , ( −1, 0, 1 )T , the Gram–Schmidt process pro-
T T
duces the orthonormal basis √1 , √2 , 0 , − 4
√ 2
, √ 5
, √ .
5 5 3 5 3 5 3 5
4.2.6. (a) √1 ( 1, 1, −1, 0 )T , √1 ( −1, 2, 1, 3 )T , √1 ( 3, −1, 2, 1 )T .

3 15 15
T
(c) Applying Gram-Schmidt to the coimage basis ( 2, 1, 0, −1 )T , 0, 12 , −1, 1
2 , gives the
orthonormal basis √1 ( 2, 1, 0, −1 )T , √1 ( 0, 1, −2, 1 )T .
6 6
T
(e) Applying Gram-Schmidt to the cokernel basis 2 1
3 , − 3 , 1, 0 , ( −4, 3, 0, 1 )T , gives the
orthonormal basis √1 ( 2, −1, 3, 0 )T , √1 ( −34, 31, 33, 14 ) . T
14 9 42
! ! ! !
1 1 1 1 1 1 1 3
4.2.7. (a) Image: √ ; kernel: √ ; coimage: √ ; cokernel: √ .
10 −3 2 1 2 −1 10 1
 
      −1
1 1 3  
1 

 1   1   1  −1 
(c) Image: √ 1, √
 4 , √  −2 ; kernel:  ;
3

42
 
14
 
2  1


−1 5 1
1
     
1 0 −1
     
1 0 1 1 1  1
coimage: √  , √  ,  ; the cokernel is {0}, so there is no basis.
    
2 1 2 0 2  1
0 1 −1
4.2.8. Applying the Gram–Schmidt process to the standard basis vectors e1 , e2 gives
       1 
√1 0 1 √
 3    , 
2 2 3
(a)  ,  ; (b)  .
0 √1 0 √2
5 3
4.2.10. (i ) (b) √1 ( 1, 1, 0 )T , √1 ( −2, 3, −5 )T , √1 ( 2, −3, −6 )T ;

5 55 66
T T T
(ii ) (b) √1 , √1 , 0 , − 21 , 0, − 21 , 0, − √1 , − √1 .
2 2 2 2

!T !T
1+ i 1− i 3− i 1+ 3i
4.2.15. (a) , , √ , √ .
2 2 2 5 2 5
!T !T
1− i 1 −1 + 2 i 3 − i 3 i
4.2.16. (a) √ , √ ,0 , √ , √ , √ .
3 3 2 6 2 6 2 6
     
.5164 −.2189 .2529
           
0. .8165 .57735  .2582 
 
 −.5200 
 
 .5454 
 
      
4.2.17. (a)  .7071 ,  −.4082 , 
     .57735 
; (c) 

.7746 
,

 
 .4926 ,
 
 
 −.2380 .
 
     
.7071 .4082 −.57735  .2582   −.5200   −.3372 
0. .4105 .6843
4.2.21. Since u1 , . . . , un form an orthonormal basis, if i < j,

(j+1) (j)
h wk , ui i = h wk , ui i,
(i)
and hence, by induction, rik = h wk , ui i = h wk , ui i. Furthermore,
(j+1) 2 (j) (j) (j)
k wk k = k wk k2 − h wk , uj i2 = k wk k2 − rjk
2
,
(i)
and so, by (4.5), k wi k2 = k wi k2 − r1i
2 2
− · · · − ri−1,i 2
= rii .
4.3.1. (a) Neither; (c) orthogonal; (f ) proper orthogonal.
4.3.3. (a) True: Using the formula (4.31) for an improper 2 × 2 orthogonal matrix,
!2 !
cos θ sin θ 1 0
= .
sin θ − cos θ 0 1
4.3.7. (Q−1 )T = (QT )T = Q = (Q−1 )−1 , proving orthogonality.
4.3.9. In general, det(Q1 Q2 ) = det Q1 det Q2 . If both determinants are +1, so is their prod-
uct. Improper times proper is improper, while improper times improper is proper.
4.3.14. False. This is true only for row interchanges or multiplication of a row by −1.
♦ 4.3.16. (a) k Q x k2 = (Q x)T Q x = xT QT Q x = xT I x = xT x = k x k2 .
4.3.18. (a) If S = ( v1 v2 . . . vn ), then S −1 = S T D, where D = diag (1/k v1 k2 , . . . , 1/k vn k2 ).

1 1 1 1 
    
1 1 1 0 −1  4 4 4 4  1 1 1 1 1
4 0 0 0
   1 1    
1 1 −1 0  − 41 − 41  1 1 −1 −1  0 1
0 0
(b)   =

4 4 
 =   4 .
  
1 −1 0 1 
 1
− 21 0 0
 1 −1 0 0 
 0 0 1
2 0
 2  1
1 −1 0 −1 1 0 0 1 −1 0 0 0 2
0 0 2 − 21
♦ 4.3.23. If S = ( v1 v2 . . . vn ), then the (i, j) entry of S T K S is viT K vj = h vi , vj i, so

S T K S = I if and only if h vi , vj i = 0 for i 6= j, while h vi , vi i = k vi k2 = 1.

    √ 
1 −3 √1 − √2 5 − √1
  5 5  5
4.3.27. (a) =    ,
2 1 √2 √1 0 √7
5 5 5
    √ 
√2 − √1 √1
√3 − √3

2 1 −1   5 30 6   5 5 5 
   q   q q 
 5 
(c) 
 0 1 3= 0 √1    0
6 2
7 15 ,

   6 6  5
 q q   q 
−1 −1 1 − √1 − 15 2 2
3 0 0 2 23
5
     

0 0 2 
0 0 1 
1 0 −1 
     
(e)  1 =  0 0 1
 0 4   0 1   4 .
     
−1 0 1 −1 0 0 0 0 2
  2     
4.3.28. 2 1 −1  √1 − 1
√ 3 0 2   
1
 3 2 3
√ 2   x 

1
 
1 2 2
  √ √     
(ii ) (a)  0 2
= 3 0 3


0
 2 −2 2 
, (b)
 y  =  −1 ;
   
     √   
2 z
2 −1 3 3 − √1 − √ 1
0 0 2 −1
2 3 2
♠ 4.3.29. 3 × 3 case:
    
4 1 0 .9701 −.2339 .0643 4.1231 1.9403 .2425
    
1 4 1  −.2571  1.9956 
  =  .2425 .9354  0 3.773 .
0 1 4 0 .2650 .9642 0 0 3.5998
♥ 4.3.32. (a) If rank A = n, then the columns w1 , . . . , wn of A are linearly independent, and so
form a basis for its image. Applying the Gram–Schmidt process converts the column basis
w1 , . . . , wn to an orthonormal basis u1 , . . . , un of img A.
   1 
1 −1  √ − 32 √ √ 
  5 
    5 5
(c) (i ) 2 3 =  √2 1  .
   3
   5  0 3
0 2 2
0 3
     
! 1 0 0 ! 1 0
−1 0   0    
4.3.34. (a) (i ) , (iii ) 
0 −1 0. (b) (i ) v = c , (iii ) v = c  0  + d 0 .
   
0 1 1
0 0 1 0 1
! !
−1.2361 .4472 .8944
4.3.35. Exercise 4.3.27: (a) b =
v , H1 = ,
1 2.0000 .8944 −.4472
! !
.4472 .8944 2.2361 −.4472
Q= , R= ;
.8944 −.4472 0 −3.1305
   
−.2361 .8944 0 −.4472
   
(c) b =  0   0 
v 1  , H1 =  0 1 ,
−1 −.4472 0 −.8944
   
0 1 0 0
   
b =  −.0954 ,
H2 = 0 .9129 .4082 
v 2    ,
.4472 0 .4082 −.9129
   
.8944 −.1826 .4082 2.2361 1.3416 −1.3416
   
Q=
 0 .9129 .4082 
, R=
 0 1.0954 2.556 
;
−.4472 −.3651 .8165 0 0 1.633

       
−1 0 0 −1 0 1 0 0
       
(e) b = 0 , H =  0 1 0 , =  0 , H2 = 0 1 0
v 1   1   vb
2    ,
−1 −1 0 0 0 0 0 1
   
0 0 −1 1 0 −1
   
Q=  0 1 0, R= 0 4 1
.
−1 0 0 0 0 −2
Exercise 4.3.29: 3 × 3 case:
   
−.1231 .9701 .2425 0
   
b =
v  1  , H1 = 
 .2425 −.9701 0,
1
0 0 0 1
   
0 1 0 0
   
b =
v  −7.411 , H2 = 0 −.9642 .2650 
2    
1 0 .2650 .9642
   
.9701 −.2339 .0643 4.1231 1.9403 .2425
   
Q=  .2425 .9354 −.2571   1.9956 
 , R=  0 3.773 ;
0 .2650 .9642 0 0 3.5998
4.4.1. (a) v2 , v4 , (c) v2 , (e) v1 .

   
1 7  

− 3   9  .7778
     
4.4.2. (a)  1 , (c)  11  ≈  1.2222 .
 3   9   
   
1 1 .1111
3 9
   
3

−1  2
   
4.4.5. Orthogonal basis:  2,  2
   ;
   
1 − 52
     
3 8  

− 1  2  15  .5333
2

 4  
 44 
  
orthogonal projection: 2
−
 2 = ≈ 2.9333 .
3
  5



 15 
 
 
5 −1.3333
1 −2 − 43
   
1 15  

−5   17  .88235
     
4.4.7. (i ) (a)  1 , (c)  19  ≈ 1.11765 .
 5   17   
   
1 1 .05882
5  17 
  23  
0  43  .5349
     
(ii ) (a)    19   .4419 
 0 , (c)  43  ≈  .
 
0 1 −.0233
− 43

♥ 4.4.9. (a) The entries of c = AT v are ci = uT i v = ui · v, and hence, by (1.11),

w = P v = A c = c1 u1 + · · · + ck uk , reproducing the projection formula (4.41).
 1 

4
 9 − 92 2
9 0
 9 − 49 2
9 

 2 8



4 4 2
 
 − 9 9 0 − 92 

(b) (ii )  − −9 , (iv )  .
 9 9   2 8 2
   0 −9 
2  9 9 
9 − 29 1
9
0 − 29 − 92 1
9
(c) P T = (A AT )T = A AT = P .
1  1
−
⊥ 3  3
4.4.12. (a) W has basis  1 , 
   0
, dim W ⊥ = 2;
0 1
 
2
 
(d) W ⊥ has basis  
 1 , dim W ⊥ = 1.
1
  1 3
3
  2 2
4.4.13. (a)  4  1 ,  0 .
 ; (b)    
−5 0 1
   
1 2
   
3 7  3  3
   
4.4.15. (a) w =  10 , z=  10 ; (c) w = −1 , z=  1 .
1 21  3  3
− 10    
10
− 13 1
3
2    
−1 6
3   3
4.4.16. (a) Span of  1 ,
 

 0; dim W ⊥ = 2. (d) Span of  ;
2 dim W ⊥ = 1.
0 1 1
Z 1
4.4.19. (a) h p , q i = p(x) q(x) dx = 0 for all q(x) = a + b x + c x2 , or, equivalently,
−1
Z 1 Z 1 Z 1
p(x) dx = x p(x) dx = x2 p(x) dx = 0. Writing p(x) = a + b x + c x2 + d x3 + e x4 ,
−1 −1 −1
2 2 2 2 2 2 2
the orthogonality conditions require 2 a + 3c+ 5 e = 0, 3 b+ 5 d = 0, 3 a+ 5 c+ 7 e = 0.
3 3 4 6 2 3 ⊥
(b) Basis: t − 5 t, t − 7t + 35 ; dim W = 2; (c) the preceding basis is orthogonal.
4.4.22. If z ∈ W2⊥ then h z , w i = 0 for every w ∈ W2 . In particular, every w ∈ W1 ⊂ W2 , and

hence z is orthogonal to every vector w ∈ W1 . Thus, z ∈ W1⊥ , proving W2⊥ ⊂ W1⊥ .
♦ 4.4.25. Suppose v ∈ (W ⊥ )⊥ . Then we write v = w + z where w ∈ W, z ∈ W ⊥ . By assumption,

for every y ∈ W ⊥ , we must have 0 = h v , y i = h w , y i + h z , y i = h z , y i. In particular,
when y = z, this implies k z k2 = 0 and hence z = 0 which proves v = w ∈ W .

4.4.29. Note: To show orthogonality of two subspaces, it suffices to check orthogonality of their
respective basis vectors.
! ! ! !
1 −2 1 2
(a) (i ) Image: ; cokernel: ; coimage: ; kernel: ;
2 1 −2 1
! ! ! !
1 −2 1 2
(ii ) · = 0; (iii ) · = 0.
2 1 −2 1
           
0 1 −3 −1 0 −3
           
(c) (i ) Image:           
 −1 ,  0 ; cokernel:  −2 ; coimage:  0 ,  1 ; kernel:  −2 ;

−2 3 1 −3 2 1
               
0 −3 1 −3 −1 −3 0 −3
               
(ii )  −1  ·  −2  =  0  ·  −2  = 0; (iii )  0 ·  −2  =  1  ·  −2  = 0.
               
−2 1 3 1 −3 1 2 1
T
2 2
4.4.30. (a) The compatibility condition is 3 b1 + b2 = 0 and so the cokernel basis is 3,1 .
(c) There are no compatibility conditions, and so the cokernel is {0}.
4.4.32. (a) Cokernel basis: ( 1, −1, 1 )T ; compatibility condition: 2 a − b + c = 0;

(c) cokernel basis: ( −3, 1, 1, 0 )T , ( 2, −5, 0, 1 )T ;
compatibility conditions: − 3 b1 + b2 + b3 = 2 b1 − 5 b2 + b4 = 0.
       
1 1
 2  2
  
1 
2
    3   
4.4.33. (a) z=
 0 ,
 w= 
 0 = −
 −2  +  −3 ;
    2







− 12 1
2 1 2
 14   3     
1 2
 17   17     
   1     
− 1    1 −1  4 
 17  1
(c) z=

,

w=
17 
 4 
=



+  .
− 4    51 
 0
 51  
3
 17   17     
5 5
− 17 17 3 3
T
1
4.4.34. (a) (i ) Fredholm requires that the cokernel basis 2, 1 be orthogonal to the right-
T
hand side ( −6, 3 ) ; (ii ) the general solution is x = −3 + 2 y with y free; (iii ) the minimum
norm solution is x = − 53 , y = 65 .
(c) (i ) Fredholm requires that the cokernel basis ( −1, 3 )T be orthogonal to the right-hand
side ( 12, 4 )T (ii ) the general solution is x = 2 + 1
2 y− 3
2 z with y, z free; (iii ) the minimum
4
norm solution is x = 7, y= − 27 , z= 6
7 .
4.4.35. If A is symmetric, ker A = ker AT = coker A, and so this is an immediate consequence of

Theorem 4.46.

4.5.1. (a) t3 = q3 (t) + 3

5 q1 (t), where
Z 1 Z 1
h t3 , q3 i 175 h t3 , q2 i 45
1= = t3 t3 − 3
5 t dt, 0= = t3 t2 − 1
3 dt,
k q3 k2 8 −1 k q2 k2 8 −1
3 h t3 , q1 i 3 Z 1
h t3 , q0 i 1 Z 1
= = t3 t dt, 0= = t3 dt.
5 k q1 k2 2 −1 k q0 k2 2 −1
5 ! d5
4.5.2. (a) q5 (t) = t5 − 10 3
9 t + 5
21 t= (t2 − 1)5 , (b) t5 = q5 (t) + 10
9 q3 (t) + 3
7 q1 (t).
10 ! dt5
s
k! dk 2k (k !)2 2
4.5.6. qk (t) = (t2 − 1)k , k qk k = .
(2 k) ! dtk (2 k) ! 2k + 1
q
♥ 4.5.10. (a) The roots of P2 (t) are ± √1 ; the roots of P3 (t) are 0, ± 3
5 .
3
4.5.11. (a) P0 (t) = 1, P1 (t) = t − 3

2 , P2 (t) = t2 − 3 t + 13
6 , P3 (t) = t3 − 9 2
2t + 33
5 t − 63
20 .
4.5.15. p0 (t) = 1, p1 (t) = t, p2 (t) = t2 − 1

3 , p3 (t) = t3 − 9
10 t.
4.5.17. L4 (t) = t4 − 16 t3 + 72 t2 − 96 t + 24, k L4 k = 24.
4.5.22. A basis for the solution set is given by ex and e2 x . The Gram-Schmidt process yields
2 (e3 − 1) x
the orthogonal basis e2 x and e .
3 (e2 − 1)

Chapter 5: Minimization and Least Squares
5.1.1. We need to minimize (3 x − 1)2 + (2 x + 1)2 = 13 x2 − 2 x + 2.

The minimum value of 25 1
13 occurs when x = 13 .
T
5.1.3. (b) ( 0, 2 )T , (d) − 32 , 3
2 .
5.1.4. Note: To minimize the distance between the point ( a, b )T to the line y = m x + c:
n o
(i ) in the ∞ norm we must minimize the scalar function f (x) = max | x − a |, | m x + c − b | ,
while (ii ) in the 1 norm we must minimize the scalar function f (x) = | x − a |+| m x + c − b |.
T
(i ) (b) all points on the line segment ( 0, y )T for 1 ≤ y ≤ 3; (d) − 23 , 3
2 .
T T
(ii ) (b) ( 0, 2 ) ; (d) all points on the line segment ( t, −t ) for −2 ≤ t ≤ −1.
5.1.7. This holds because the two triangles in the figure are congruent. According to Exercise
5.1.6(c), when k a k = k b k = 1, the distance is | sin θ | where θ is the angle between a, b, as
ilustrated:
b
a
θ
| a x0 + b y0 + c z 0 + d | 1
♥ 5.1.9. (a) The distance is given by √ . (b) √ .
a 2 + b 2 + c2 14
1 1
5.2.1. x = 2, y= = −2, with f (x, y, z) = − 32 . This is the global minimum because the
2 ,z 
1 1 0
 
coefficient matrix  
 1 3 1  is positive definite.
0 1 1
5.2.3. (b) Minimizer: x = 29 , y = 29 ; minimum value: 32

9 .
(d) Minimizer: x = − 21 , y = −1, z = 1; minimum value: − 54 .
(f ) No minimum.
T
5.2.5. (a) p(x) = 4 x2 −24 x y+45 y 2 +x−4 y+3; minimizer: x⋆ = 1 1
24 , 18 ≈ ( .0417, .0556 )T ;
minimum value: p(x⋆ ) = 419
144 ≈ 2.9097.
2 2
(b) p(x) = 3 x + 4 x y + y − 8 x − 2 y; no minimizer since K is not positive definite.
36
T
5.2.7. (a) maximizer: x⋆ = 10 3
11 , 11 ; maximum value: p(x⋆ ) = 16
11 .
♦ 5.2.9. Let x⋆ = K −1 f be the minimizer. When c = 0, according to the third expression in

(5.14), p(x⋆ ) = − (x⋆ )T K x⋆ ≤ 0 because K is positive definite. The minimum value is 0 if
and only if x⋆ = 0, which occurs if and only if f = 0.
5.2.13. False. See Example 4.51 for a counterexample.
T
5.3.1. Closest point: 6 38 36
7 , 35 , 35 ≈ ( .85714, 1.08571, 1.02857 )T ; distance: √1 ≈ .16903.
35
5.3.2. (a) Closest point: ( .8343, 1.0497, 1.0221 )T ; distance: .2575.

q
7 7 7 7 T 11
5.3.4. (a) Closest point: 4, 4, 4, 4 ; distance: 4 .
T
(c) Closest point: ( 3, 1, 2, 0 ) ; distance: 1.
q
3 3 3 3 T 7
5.3.6. (i ) Exercise 5.3.4: (a) Closest point: 2, 2, 2, 2 ; distance: 4.
T
(c) Closest point: ( 3, 1, 2, 0 ) ; distance: 1.
T
(ii ) Exercise 5.3.4: (a) Closest point: 25 25 25 25
14 , 14 , 14 , 14 ≈ ( 1.7857, 1.7857, 1.7857, 1.7857 )T ;
q
215
distance: 14 ≈ 3.9188.
T
(c) Closest point: 28 11 16
9 , 9 , 9 ,0 ≈ ( 3.1111, 1.2222, 1.7778, 0 )T ;
q
32
distance: 9 ≈ 1.8856.
q
8
5.3.8. (a) 3.
T
1 1
5.3.12. 2, −2, 2 .
T
5.3.14. Orthogonal basis: ( 1, 0, 2, 1 )T , ( 1, 1, 0, −1 )T , 1
2 , −1, 0, − 2
1
;
T
closest point = orthogonal projection = − 32 , 2, 32 , 43 .
  !
8
1  5 
1.6
5.4.1. (a) 2, (b) = .
28 .4308
65
1 8
5.4.2. (b) x = − 25 , y = − 21 ; (d) x = 13 , y = 2, z = 43 .
5.4.5. The solution is x⋆ = ( −1, 2, 3 )T . The least squares error is 0 because b ∈ img A and so
x⋆ is an exact solution.
5.4.8. The solutions are, of course, the same:
 
.8 −.43644
  ! !
 .4 .65465  5 0 −.04000
(b) Q = 

,
 R= , x= ;
 .2 −.43644  0 4.58258 −.38095
.4 .43644

 
.18257 .36515 .12910  
  5.47723 −2.19089 0


.36515 −.18257 .90370 
,



(d) Q =  R= 0 1.09545 −3.65148 
,
 0 .91287 .12910 


0 0 2.58199
−.91287 0 .38730
x = ( .33333, 2.00000, .75000 )T .
T T
5.4.10. (a) − 17 , 0 , (b) 9 4
14 , 31 .
5.5.1. (b) y = 1.9 − 1.1 t.
5.5.3. (a) y = 3.9227 t − 7717.7; (b) $147, 359 and $166, 973.
5.5.6. (a) The least squares exponential is y = e4.6051−.1903 t and, at t = 10, y = 14.9059.
(b) Solving e4.6051−.1903 t = .01, we find t = 48.3897 ≈ 49 days.
5.5.8. (a) .29 + .36 t, (c) − 1.2308 + 1.9444 t.
♦ 5.5.12.
1 m
X 1 m
X 2t m
X ( t )2 m
X
(ti − t )2 = t2i − ti + 1 = t2 − 2 ( t )2 + ( t )2 = t2 − ( t )2 .
m i=1 m i=1 m i=1 m i=1
10
5.5.13. (a) y = 2 t − 7; 5
2 4 6 8 10
-5
14
12
10
2
(b) y = t + 3 t + 6; 8
6
4
2
-3 -2 -1 1 2
5
4
3
3 2
(d) y = − t + 2t − 1. 2
1
-1.5 -1 -0.5 0.5 1 1.5 2 2.5

-1
-2
-3
5.5.14. (a) p(t) = − 15 (t − 2) + (t + 3) = 17

5 + 4
5 t,
1 1 1 5 1 2
(b) p(t) = 3 (t − 1)(t − 3) − 4 t(t − 3) + 24 t(t − 1) = 1 − 8 t+ 8t ,
(d) p(t) = 1
2 t (t − 2)(t − 3) − 2 t (t − 1)(t − 3) + 3
2 t (t − 1)(t − 2) = t2 .

5.5.17. The quadratic least squares polynomial is y = 4480.5 + 6.05 t − 1.825 t2 , and y = 1500 at
t = 42.1038 seconds.
1 2
5.5.20. (a) p2 (t) = 1 + t + 2t , whose maximal error over [ 0, 1 ] is .218282.
5.5.22. p(t) = .9409 t + .4566 t2 − .7732 t3 + .9330 t4 . The graphs are very close over the interval
0 ≤ t ≤ 1; the maximum error is .005144 at t = .91916. The functions rapidly diverge above
1, with tan t → ∞ as t → 12 π, whereas p( 12 π) = 5.2882. The first graph is on the interval
1
[ 0, 1 ] and the second on [ 0, 2 π ]:
14
1.5
12
1.25
10
1
8
0.75
6
0.5
4
0.25 2
0.2 0.4 0.6 0.8 1 0.2 0.4 0.6 0.8 1 1.2 1.4
5.5.26. (a) p1 (t) = 14 + 7

2 t, p2 (t) = p3 (t) = 14 + 7
2 t+ 1
14 (t2 − 2).

5.5.28. (a) p4 (t) = 14 + 7
2 t+ 1
14 (t2 − 2) − 5
12 t4 − 31
7 + 72
35 .
t3 − t t2
♦ 5.5.31. q0 (t) = 1, q1 (t) = t − t, q2 (t) = t2 − t − t − t2 ,
t2 − t2
t3 − t t2
q0 = t0 , q1 = t1 − t t0 , q2 = t2 − t1 − t − t2 t0 ,
t2 − t2
2
2 2 2 2
2 t3 − t t2
k q0k = 1, k q1k = t2 − t , k q2k = t4 − t2 − .
t2 − t2
5.5.35. (a) For example, an interpolating polynomial for the data (0, 0), (1, 1), (2, 2) is the straight
line y = t.
♦ 5.5.38. (a) If p(xk ) = a0 + a1 xk + a2 x2k + · · · + an xn

k = 0 for k = 1, . . . , n + 1, then V a = 0
where V is the (n + 1) × (n + 1) Vandermonde matrix with entries vij = xi−1
j for
i, j = 1, . . . , n + 1. According to Lemma 5.16, if the sample points are distinct, then V is a
nonsingular matrix, and hence the only solution to the homogeneous linear system is a = 0,
which implies p(x) ≡ 0.
f (x + h) − f (x − h) f (x + h) − 2 f (x) + f (x − h)
♥ 5.5.39. (a) f ′ (x) ≈ ; (b) f ′′ (x) ≈
2h h2
Z b h i
1
♥ 5.5.40. (a) Trapezoid Rule: f (x) dx ≈ 2 (b − a) f (x0 ) + f (x1 ) .
a
Z b h i
1
(b) Simpson’s Rule: f (x) dx ≈ 6 (b − a) f (x0 ) + 4 f (x1 ) + f (x2 ) .
a

   
1 0 1
   
5.5.41. The sample matrix is A =  0 1  .5 
 ; the least squares solution to A x = y =  
−1 0 .25
3 1
gives g(t) = 8 cos π t + 2 sin π t.
5.5.43. (a) g(t) = .538642 et − .004497 e2 t , (b) .735894.

(c) The maximal error is .745159 which occurs at t = 3.66351. 1
0.8
♥ 5.5.45. (a) n = 1, k = 4: p(t) = .4172 + .4540 cos t; 0.6
maximal error: .1722; 0.4
0.2
-3 -2 -1 1 2 3
3 9 9 9 9 2
5.5.47. (a) 7 + 14 t; (b) 28 + 7 t− 14 t .
♥ 5.5.51. (a) 1.875 x2 − .875 x, (c) 1.7857 x2 − 1.0714 x + .1071.

3 15 25 2 2 5
5.5.54. (i ) 28 − 14 t + 14 t ≈ .10714 − 1.07143 t + 1.78571 t ; maximal error: 28 = .178571 at
2 25 50 2 2 2
t = 1; (ii ) 7 − 14 t + 21 t ≈ .28571 − 1.78571 t + 2.38095 t ; maximal error: 7 = .285714
at t = 0.
1 9
5.5.56. (a) z = x + y − 3 , (b) z = 10 (x − y).

1 4 3 2 6 2
5.5.58. (a) 5 + 7 − 12 + 2t
3
= − 35 + 7t .
5.5.60. 1.718282 + .845155 (2 t − 1) + .139864 (6 t2 − 6 t + 1) + .013931 (20 t3 − 30 t2 + 12 t − 1)

= .99906 + 1.0183 t + .421246 t2 + .278625 t3 .
♠ 5.5.63. .459698 + .427919 (2 t − 1) − .0392436 (6 t2 − 6 t + 1) − .00721219 (20 t3 − 30 t2 + 12 t − 1)

= −.000252739 + 1.00475 t − .0190961 t2 − .144244 t3 .
0.5

 −1.25 (x + 1)3 + 4.25 (x + 1) − 2, −1 ≤ x ≤ 0, -1 -0.5 0.5 1
5.5.67. (a) u(x) =  -0.5
1.25 x3 − 3.75 x2 + .5 x − 1, 0 ≤ x ≤ 1. -1
-1.5
-2
3
2.5

 2
 3 (x − 1)3 − 11
3 (x − 1) + 3, 1 ≤ x ≤ 2, 1.5
2
(c) u(x) = 
 − 1 (x − 2)3 + 2 (x − 2)2 − 5
1
3 3 (x − 2), 2 ≤ x ≤ 4. 0.5
1.5 2 2.5 3 3.5 4

5.5.68. 1
0.5

 −5.25 (x + 1)3 + 8.25 (x + 1)2 − 2, −1 ≤ x ≤ 0, -1 -0.5 0.5 1
(a) u(x) =  -0.5
4.75 x3 − 7.5 x2 + .75 x + 1, 0 ≤ x ≤ 1. -1
-1.5
-2
3
2.5

 3.5 (x − 1)3 − 6.5 (x − 1)2 + 3,
2
 1 ≤ x ≤ 2, 1.5
(c) u(x) =  1
 − 1.125 (x − 2)3 + 4 (x − 2)2 − 2.5 (x − 2), 2 ≤ x ≤ 4. 0.5
1.5 2 2.5 3 3.5 4

-0.5
1



 x3 − 2 x2 + 1, 0 ≤ x ≤ 1, 0.8

 0.6
5.5.69. (a) u(x) =  (x − 1)2 − (x − 1), 1 ≤ x ≤ 2, 0.4

 0.2


− (x − 2)3 + (x − 2)2 + (x − 2), 2 ≤ x ≤ 3. -0.2
0.5 1 1.5 2 2.5 3
♥ 5.5.75. (a)



 1− 19
15 x+ 4
15 x3 , 0 ≤ x ≤ 1, 1




0.8
7
C0 (x) =  − 15 (x − 1) + 54 (x − 1)2 − 1
(x − 1)3 ,
0.6
3 1 ≤ x ≤ 2,


0.4



 2 1 2 1 3
0.2
15 (x − 2) − 5 (x − 2) + 15 (x − 2) , 2 ≤ x ≤ 3, 0.5 1 1.5 2 2.5 3
 8


 5 x− 3
5 x3 , 0 ≤ x ≤ 1, 1




0.8
1
(x − 1) − 95 (x − 1)2 + (x − 1)3 ,
0.6
C1 (x) =  1 − 5 1 ≤ x ≤ 2, 0.4



 0.2


− 54 (x − 2) + 6
5 (x − 2)2 − 2
5 (x − 2)3 , 2 ≤ x ≤ 3, 0.5 1 1.5 2 2.5 3



 − 25 x + 2
5 x3 , 0 ≤ x ≤ 1, 1




0.8
4 6 2 3 0.6
C2 (x) =  5 (x − 1) + 5 (x − 1) − (x − 1) , 1 ≤ x ≤ 2, 0.4



 0.2

 1 9 2 3 3
5 (x − 2) − 5 (x − 2) + 5 (x − 2) , 2 ≤ x ≤ 3, 0.5 1 1.5 2 2.5 3

1 1 3
15 x − 15 x , 0 ≤ x ≤ 1,

 1



 0.8

C3 (x) = − 12 15 (x − 1) − 15 (x − 1)2 + 31 (x − 1)3 ,
0.6
 1 ≤ x ≤ 2,


0.4


 7
 4 2 4 3
0.2
15 (x − 2) + 5 (x − 2) − 15 (x − 2) , 2 ≤ x ≤ 3. 0.5 1 1.5 2 2.5 3
(b) It suffices to note that any linear combination of natural splines is a natural spline. More-
over, u(xj ) = y0 C0 (xj ) + y1 C1 (xj ) + · · · + yn Cn (xj ) = yj , as desired.

−ix
5.6.1. (a) (i ) c0 = 0, c1 = − 12 i , c2 = c−2 = 0, c3 = c−1 = 1 1
2 i , (ii ) 2 i e − 12 i e i x = sin x;
√ √ √
(c) (i ) c0 = 13 , c1 = 3−123 i , c2 = 1−123 i , c3 = c−3 = 0, c4 = c−2 = 1+123 i ,
√ √ √ √ √
c5 = c−1 = 3+123 i , (ii ) 1+123 i e− 2 i x + 3+123 i e− i x + 13 + 3−123 i e i x + 1−123 i e2 i x
= 13 + 12 cos x + √ 1
sin x + 61 cos 2 x + √ 1
sin 2 x.
2 3 2 3
5.6.2. (a) (i ) f0 = 2, f1 = −1, f2 = −1. (ii ) e− i x + e i x = 2 cos x;

(c) (i ) f0 = 6, f1 = 2 + 2 e2 π i /5 + 2 e− 4 π i /5 = 1 + .7265 i ,
f2 = 2 + 2 e2 π i /5 + 2 e4 π i /5 = 1 + 3.0777 i ,
f3 = 2 + 2 e− 2 π i /5 + 2 e− 4 π i /5 = 1 − 3.0777 i ,
f4 = 2 + 2 e− 2 π i /5 + 2 e4 π i /5 = 1 − .7265 i ;
(ii ) 2 e− 2 i x + 2 + 2 e i x = 2 + 2 cos x + 2 i sin x + 2 cos 2 x − 2 i sin 2 x.
6 6
6
5 5 5
4 4 4
♠ 5.6.3. 3 3 3
2 2 2
1 1 1
1 2 3 4 5 6 1 2 3 4 5 6
1 2 3 4 5 6
The interpolants are accurate along most of the interval, but there is a noticeable problem
near the endpoints x = 0, 2 π. (In Fourier theory, [ 19, 61 ], this is known as the Gibbs phe-
nomenon.)
40 40 40
30 30 30
♠ 5.6.4. (a) 20 20 20
10 10 10
1 2 3 4 5 6 1 2 3 4 5 6
1 2 3 4 5 6
1 1 1
0.5 0.5 0.5
(c) 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6
-0.5 -0.5 -0.5
-1 -1 -1
♦ 5.6.7. (a) (i ) i , − i ; (ii ) e2 πk i /5 for k = 1, 2, 3 or 4; (iii ) e2 πk i /9 for k = 1, 2, 4, 5, 7 or 8.
2 2
2
1.5 1.5 1.5
1 1 1
♠ 5.6.10. 0.5 0.5 0.5
1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6
-0.5 -0.5 -0.5
-1 -1 -1
Original function, 11 mode compression, 21 mode compression.
The average absolute errors are .018565 and .007981; the maximal errors are .08956 and
.04836, so the 21 mode compression is about twice as accurate.

♣ 5.6.13. Very few are needed. In fact, if you take too many modes, you do worse! For example,
if ε = .1,
2 2 2 2 2
1.5 1.5 1.5 1.5 1.5
1 1 1 1 1
0.5 0.5 0.5 0.5 0.5
1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6
plots the noisy signal and the effect of retaining 2 l + 1 = 3, 5, 11, 21 modes. Only the first
three give reasonable results. When ε = .5 the effect is even more pronounced:
2.5 2.5 2.5 2.5
2
2 2 2 2
1.5
1.5 1.5 1.5 1.5
1 1 1 1 1
0.5 0.5 0.5 0.5 0.5
1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6
-0.5 -0.5 -0.5 -0.5
   1     3 
0 0
   2     4 
1  1    1 
  −  1 − + 1 i 
♠ 5.6.17. (a) f =
(0)
c = 2
 ,
 
(1)
c =  2

, c = c

 
(2)
 ,
1
= 4

4 
;

1
 
 1 
 
 
2


− 14 

3 1 3 1 1
2 −2 2 −4 + 4 i
    
   1 
2π
1 1

π  
π   2
π   2
π   √ 
         2+1 
 3π  0   1π   1π   √ π
4     2   4   8 2 
         
1  1   1     
 π  π      0 
2  2   2 π   0  



         √ 
1  1     1   2−1
 π
4  (0)
 π
2  (1)
 0 
  (2)

 4 π 
 (3)  √ π

 8 2 .
(c) f =  
, c

= 
3 
, c = 
 1
, c

= 
 1
, c = c

=  
 0   π  
 2π   2π  0
     
  4   √ 
         
1  1   1   1+ i   2−1
 4π
 
 4π
 
 4π 
 
 8 π
   √ π

 8 2 
         
 1π  1π  1π   0   
2 
 
4 
 
 2






  0 
 √ 
3 3 1 1− i 2+1
4 π 4 π − 4 π 8 π √ π
        8 2
1 1 2 0
       
 −1   1  0 0
♠ 5.6.18. (a) c = 

,
 f (0) =

,
 f (1) =

,
 f = f (2) =
 

 1  −1   −2  4
−1 −1 0 0

Chapter 6: Equilibrium
!   !
18
3 −2  5 
3.6
6.1.1. (a) K = ; (b) u = = ; (c) the first mass has moved the
−2 3 17 3.4
5
T
farthest; (d) e = 18 1 17
5 , − 5, − 5 = ( 3.6, − .2, − 3.4 )T , so the first spring has stretched
the most, while the third spring experiences the most compression.
!   !
3 −2 7 7.0
6.1.3. Exercise 6.1.1: (a) K = ; (b) u =   = ;
−2 2 17 8.5
2
(c) the second mass has moved the farthest;
T
(d) e = 7, 3
2 = ( 7.0, 1.5 )T , so the first spring has stretched the most.
♣ 6.1.7. Top and bottom support; constant force:
12
0.4
10
8 0.2
6
20 40 60 80 100
4 -0.2
2
-0.4
20 40 60 80 100
Top support only; constant force:
50 1
40 0.8
30 0.6
20 0.4
10 0.2
20 40 60 80 100 20 40 60 80 100
6.1.8. (a) For maximum displacement of the bottom mass, the springs should be arranged from
weakest at the top to strongest at the bottom, so c1 = c = 1, c2 = c′ = 2, c3 = c′′ = 3.
6.1.13. Denoting the gravitation force by g:

    
2 −1 0 u g
1   1   
(a) p(u) = ( u1 u2 u3 )  −1 2 −1     
  u2  − ( u1 u2 u3 ) g 
2
0 −1 1 u3 g
= u21 − u1 u2 + u22 − u2 u3 + 1
2 u23 − g (u1 + u2 + u3 ).
! ! !
1 3 −2 u1 4
6.1.14. (a) p(u) = ( u1 u2 ) − ( u1 u2 )
2 −2 3 u2 3

= 3
2 u21 − 2 u1 u2 + 3
2 u22 − 4 u1 − 3 u2 , so p(u⋆ ) = p 3.6, 3.4 = − 12.3.
(b) For instance, p(1, 0) = −2.5, p(0, 1) = −1.5, p(3, 3) = −12.
44
6.1.16. (a) Two masses, both ends fixed, c1 = 2, c2 = 4, c3 = 2, f = ( −1, 3 )T ;

equilibrium: u⋆ = ( .3, .7 )T .
(c) Three masses, top end fixed, c1 = 1, c2 = 3, c3 = 5, f = ( 1, 1, −1 )T ;
T
equilibrium: u⋆ = − 21 , − 65 = ( −.5, −.8333 )T .
6.2.1. (b) (d)
 
1 −1 0 0
      
1 0 −1 0 3 −1 −1 u 3


   1   
6.2.2. (a) A = 1

0 0 −1 
;

(b)  −1
 3 −1   u2  =  0 
   
.

0 1 −1 0 −1 −1 2 u3 0
0 1 0 −1
T
(c) u = 15 9 3
8 , 8, 2 = ( 1.875, 1.125, 1.5 )T ;
T
(d) y = v = A u = 3 3 15
4, 8, 8 , − 83 , 9
8 = ( .75, .375, 1.875, − .375, 1.125 )T .
(e) The bulb will be brightest when connected to wire 3, which has the most current
flowing through.
♠ 6.2.6. None.
♠ 6.2.8. (a) The potentials remain the same, but the currents are all twice as large.
6.2.12. (a) True, since they satisfy the same systems of equilibrium equations K u = − AT C b = f .
(b) False, because the currents with the batteries are, by (6.37), y = C v = C A u + C b,
while for the current sources they are y = C v = C A u.
6.2.13. (a) (i ) u = ( 2, 1, 1, 0 )T , y = ( 1, 0, 1 )T ; (iii ) u = ( 3, 2, 1, 1, 1, 0 )T , y = ( 1, 1, 0, 0, 1 )T .

T T T T
3 1
6.2.14. (i ) u = 2 , 2 , 0, 0 , y = 1, 12 , 1
2 ; (iii ) u = 7 4 1
3 , 3 , 3 , 0, 0, 0 , y = 1, 1, 13 , 13 , 1
3 .
6.2.17. (a) If f are the current sources at the nodes and b the battery terms, then the nodal
voltage potentials satisfy AT C A u = f − AT C b.
(b) By linearity, the combined potentials (currents) are obtained by adding the potentials
(currents) due to the batteries and those resulting from the current sources.
(c) 1
2 P = p(u) = 1
2 uT K u − uT (f − AT C b).
6.3.1. 8 cm

6.3.3. (a) For a unit horizontal force on the two nodes, the displacement vector is
u = ( 1.5, − .5, 2.5, 2.5 )T , so the left node has moved slightly down and three times as far
to the right, while the right node has moved five times as far up and to the right. Note that
the force on the left node is transmitted through the top bar to the right node, which ex-
plains why it moves significantly further. The stresses are e = ( .7071, 1, 0, − 1.5811 )T , so
the left and the top bar are elongated, the right bar is stress-free, and the reinforcing bar is
significantly compressed.
  3 1
0 1 0 0 2 u1 − 2 v1 − u2 = f1 ,
 

 −1 0 1 0 
 1 3
  − u1 +
2 2 v1 = g1 ,
♥ 6.3.5. (a) A = 

0 0 0 1 ;
 (b)

 0 0 √1 √ 
1
 − u1 + 32 u2 + 1
2 v2 = f2 ,
 2 2
− √1 √1 0 0 1 3
2 2 2 u2 + 2 v2 = g2 .
(c) Stable, statically indeterminate.

3! !
−1 2 ,
(d) Write down f = K e1 , so f1 = f2 =
. The horizontal bar is under the
0 − 12
most stress: it is compressed by −1; the upper left to lower right bar is compressed − √1 ,
2
while all other bars are stress free.
 
0 1 0 0 0 0
 
 

 − √3 − √1 √3 √1 0 0 

 10 10 10 10 
 
♥ 6.3.8. (a) A =  −1 0 0 0 1 0 
 
 
 

 0 0 − √3 √1 √3 − √1 
 10 10 10 10 

0 0 0 0 0 1
 
0 1 0 0 0 0
 
 − .9487 − .3162 .9487 .3162 0 0 
 
 
= 
−1 0 0 0 1 0 ;


 0 0 − .9487 .3162 .9487 − .3162  
0 0 0 0 0 1
(b) One instability: the mechanism of simultaneous horizontal motion of the three nodes.
(c) No net horizontal force: f1 +f2 +f3 = 0. For example, if f1 = f2 = f3 = ( 0, 1 )T , then e =
q q T
3
2,
5 3
2, −2,
5 3
2, 2 = ( 1.5, 1.5811, −1.5, 1.5811, 1.5 )T , so the compressed diagonal bars
have slightly more stress than the compressed vertical bars or the elongated horizontal bar.
(d) To stabilize, add in one more bar starting at one of the fixed nodes and going to one of
the two movable nodes not already connected to it.
q q T
(e) In every case, e = 3
2,
5 3
2, −2,
5 3
2, 2,0 = ( 1.5, 1.5811, −1.5, 1.5811, 1.5, 0 )T , so
the stresses on the previous bars are all the same, while the reinforcing bar experiences no
stress. (See Exercise 6.3.19 for the general principle.)

♣ 6.3.11. (a)
 
−1 0 0 1 0 0 0 0 0 0 0 0
 
 0 −1 0 0 0 0 0 1 0 0 0 0 
 
 √1 
 0 0 −1 0 0 0 0 0 0 0 0 
 2 
A=
 0 0 0 √1 − √1 0 − √1 0 0 0 0 1 
;
 2 2 2 
 
 √1 √1 √1 √1 
 0 0 0 0 − 0 0 0 − 0 
 2 2 2 2
0 0 0 0 0 0 0 √1 − √1 0 − √1 √1
2 2 2 2
           
1 0 0 0 0 0
           
0 1 0  0  0  0
           
           
0 0 1  0  0  0
           
           
1 0 0  0  0  0
           
0 1 0  0  0  −1 
           
           
0 0 1  0  −1   0
           
(b) v1 =  ,
1
v2 =  ,
0
v3 =  ,
0
v4 = 
 0
, v5 = 
 0
, v6 = 
 1
;
           
           
0 1 0  0  0  0
           
           
0 0 1  −1   0  0
           
           
1 0 0  0  1  0
           
           
0 1 0  1  0  0
0 0 1 0 0 0
(c) v1 , v2 , v3 correspond to translations in, respectively, the x, y, z directions;

(d) v4 , v5 , v6 correspond to rotations around, respectively, the x, y, z coordinate axes;
 
1 0 0 −1 0 0 0 0 0 0 0 0
 
 0 1 0 0 0 0 0 −1 0 0 0 0
 

 0

0 1 0 0 0 0 0 0 0 0 −1 



 −1 0 0 2 − 12 − 21 − 21 1
2 0 − 12 0 1
2
 
 0
 0 0 − 21 1
2 0 1
2
1
−2 0 0 0 0
 
 0
 0 0 − 21 0 1
2 0 0 0 1
2 0 − 21 

(e) K =  ;
 0
 0 0 − 21 1
2 0 2
1
− 21 0 0 0 0
 1 1
 0
 −1 0 2 − 12 0 − 21 2 − 21 0 − 12 2

 1
 0
 0 0 0 0 0 0 − 12 2
1
0 1
2 −2  

 0

0 0 − 21 0 1
2 0 0 0 1
2 0 − 21 



 0 0 0 0 0 0 0 − 12 1
2 0 1
2 − 21 

1
0 0 −1 2 0 − 12 0 1
2
1
−2 − 12 − 21 2
(f ) For fi = ( fi , gi , hi )T we require f1 + f2 + f3 + f4 = 0, g1 + g2 + g3 + g4 = 0,
h1 + h2 + h3 + h4 = 0, h3 = g4 , h2 = f4 , g2 = f3 , i.e., there is no net horizontal force and
no net moment of force around any axis.
(g) You need to fix three nodes. Fixing two still leaves a rotation motion around the line
connecting them.
(h) Displacement of the top node: u4 = ( −1, −1, −1 )T ; since e = ( −1, 0, 0, 0 )T , only the
vertical bar experiences compression of magnitude 1.

6.3.14. (a) 3 n.
(b) Example: a triangle each of whose nodes is connected
to the ground by two additional, non-parallel bars.
♦ 6.3.18. (a) We are assuming that f ∈ img K = coimg A = img AT , cf. Exercise 3.4.32. Thus, we
can write f = AT h = AT C g where g = C −1 h.
(b) The equilibrium equations K u = f are AT C A u = AT C g which are the normal equa-
tions (5.36) for the weighted least squares solution to A u = g.
 
√1 √1 0 0 0
 2 2 
 
♥ 6.3.20. (a) A⋆ = 
 −1 0 1 0 0;
 
0 0 − √1 √1 √1
2 2 2
 3 1 
2 2 −1 0 0
 
 1 1 
 0 0 0
 2 2 
 
 
K ⋆ u = f ⋆ where K ⋆ = 
 −1 0 3
2 − 21 − 12 

.
 
 1


0 0 − 21 1
2 2

0 0 − 21 1
2
1
2
(b) Unstable, since there are two mechanisms prescribed by the kernel basis elements
( 1, −1, 1, 1, 0 )T , which represents the same mechanism as when the end is fixed, and
( 1, −1, 1, 0, 1 )T , in which the roller and the right hand node move horizontally to the right,
while the left node moves down and to the right.
6.3.24. (a) True. Since K u = f , if f 6= 0 then u 6= 0 also.

Chapter 7: Linearity
7.1.1. (b) Not linear; (d) linear (f ) not linear.

! !
2 0
7.1.2. (a) F (0, 0) = 6= , (c) F (− x, − y) = F (x, y) 6= − F (x, y).
0 0
7.1.3. (b) Not linear; (d) linear; (f ) linear.

 
   

1 0 0 0 0 1 1 0 0
 √     
1
7.1.5. (b) 0
 − 23 
, (d) 1
 0 0, (f ) 0
 1 0.
 √2 
3 1 0 1 0 0 0 0
0 2 2
7.1.9. No, because linearity would require

         
0 1 1 1 1
         
L
 1 = L   −1  −  −1   = L 0  − L −1  = 3 6= −2.
       
−1 0 0 −1 0
7.1.15. (a) L[ c X + d Y ] = A(c X + d Y ) = c A X + d A Y = c L[ X ] + d L[ Y ];

 
a 0 b 0
 
0 a 0 b
matrix representative: 

.
c 0 d 0
0 c 0 d
7.1.16. (b) Not linear; codomain space = Mn×n . (d) Not linear; codomain space = Mn×n .
(f ) Linear; codomain space = R.
7.1.19. (b) Not linear; codomain = R. (d) Linear; codomain = R.

(f ) Linear; codomain = C (R). (h) Linear; codomain = C0 (R).
1
Z bh i
7.1.22. Iw [ c f + d g ] = c f (x) + d g(x) w(x) dx
a
Z b Z b
=c f (x) w(x) dx + d g(x) w(x) dx = c Iw [ f ] + d Iw [ g ].
a a
∂2 h i ∂2 h i
7.1.24. ∆[ c f + d g ] = 2
c f (x, y) + d g(x, y) + 2
c f (x, y) + d g(x, y)
∂x ∂y
   
∂2f ∂2f ∂2g ∂2g
= c + +d +  = c ∆[ f ] + d ∆[ g ].
∂x2 ∂y 2 ∂x2 ∂y 2
7.1.26. (a) Gradient: ∇(c f + d g) = c ∇f + d ∇g; domain is space of continuously differentiable

scalar functions; codomain is space of continuous vector fields.
49
! ! ! !
1 0 0 1 0 0 0 0
7.1.27. (b) dimension = 4; basis: , , , .
0 0 0 0 1 0 0 1
(d) dimension = 4; basis given by L0 , L1 , L2 , L3 , where Li [ a3 x3 + a2 x2 + a1 x + a0 ] = ai .
! !
0 1 0 0
7.1.28. True. The dimension is 2, with basis , .
0 0 0 1
T
7.1.30. (a) a = ( 3, −1, 2 )T , (c) a = 5 1 5
4, −2, 4 .
7.1.34. (a) 9 − 36 x + 30 x2 , (c) 1.
7.1.37. (a) S ◦ T = T ◦ S = clockwise rotation by 60◦ = counterclockwise rotation by 300◦ ;

(b) S ◦ T = T ◦ S = reflection in the line y = x; (c) S ◦ T = T ◦ S = rotation by 180◦ ;
(e) S ◦ T = T ◦ S = O.
   
1 0 0 0 −1 0
   
7.1.39. (a) R = 
0 0 −1 
, S = 1
 0 0;
0 1 0 0 0 1
   
−10 0 0 0 1
   
(b) R ◦ S = 0 0 −1  =
6 S ◦ R =  1 0 0 .
   
1 0 0 0 1 0
Under R ◦ S, the basis vectors e1 , e2 , e3 go to e3 , − e1 , − e2 , respectively.
Under S ◦ R, they go to e2 , e3 , e1 .
7.1.41. (a) L = E ◦ D where D[ f (x) ] = f ′ (x), E[ g(x) ] = g(0). No, they do not commute:
D ◦ E is not even defined since the codomain of E, namely R, is not the domain of D, the
space of differentiable functions. (b) e = 0 is the only condition.
7.1.43. Given L = an Dn + · · · + a1 D + a0 , M = bn Dn + · · · + b1 D + b0 , with ai , bi constant,

the linear combination c L + d M = (c an + d bn ) Dn + · · · + (c a1 + d b1 ) D + (c a0 + d b0 ),
is also a constant coefficient linear differential operator, proving that it is a subspace of the
space of all linear operators. A basis is Dn , Dn−1 , . . . , D, 1 and so its dimension is n + 1.
X X
7.1.46. If p(x, y) = cij xi y j then p(x, y) = cij ∂xi ∂yj is a linear combination of linear opera-
tors, which can be built up as compositions ∂xi ◦ ∂yj = ∂x ◦ · · · ◦ ∂x ◦ ∂y ◦ · · · ◦ ∂y of the basic
first order linear partial differential operators.
7.1.51. (a) The inverse is the scaling transformation that halves the length of each vector.
(b) The inverse is counterclockwise rotation by 45◦ . (d) No inverse.
   
√1 √1 √1 − √1
 2 2  2 2
7.1.52. (b) Function:  ; inverse:  .
− √1 √1 √1 √1
2 2 2 2
 
1 1
(d) Function: 2 2
; no inverse.
1 1
2 2

♦ 7.1.55. If L ◦ M = L ◦ N = I W , M ◦ L = N ◦ L = I V , then, by associativity,

M = M ◦ I W = M ◦ (L ◦ N) = (M ◦ L) ◦ N = I V ◦ N = N.
♥ 7.1.58. (a) Every vector in V can be uniquely written as a linear combination of the basis ele-
ments: v = c1 v1 + · · · + cn vn . Assuming linearity, we compute
L[ v ] = L[ c1 v1 + · · · + cn vn ] = c1 L[ v1 ] + · · · + cn L[ vn ] = c1 w1 + · · · + cn wn .
Since the coefficients c1 , . . . , cn of v are uniquely determined, this formula serves to uniquely
define the function L : V → W . We must then check that the resulting function is linear.
Given any two vectors v = c1 v1 + · · · + cn vn , w = d1 v1 + · · · + dn vn in V , we have
L[ v ] = c1 w1 + · · · + cn wn , L[ w ] = d1 w1 + · · · + dn wn .
Then, for any a, b ∈ R,
L[ a v + b w ] = L[ (a c1 + b d1 ) v1 + · · · + (a cn + b dn ) vn ]
= (a c1 + b d1 ) w1 + · · · + (a cn + b dn ) wn

= a c1 w1 + · · · + cn wn + b d1 w1 + · · · + dn wn = a L[ v ] + d L[ w ],
proving linearity of L.
(b) The inverse is uniquely defined by the requirement that L−1 [ wi ] = vi , i = 1, . . . , n.
Note that L ◦ L−1 [ wi ] = L[ vi ] = wi , and hence L ◦ L−1 = I W since w1 , . . . , wn is a basis.
Similarly, L−1 ◦ L[ vi ] = L−1 [ wi ] = vi , and so L−1 ◦ L = I V .
(c) If A = ( v1 v2 . . . vn ), B = ( w1 w2 . . . wn ), then L has matrix representative B A−1 ,
while L−1 has matrix representative A B −1 .
! !
3 5 −1 2 −5
(d) (i ) L = , L = .
1 2 −1 3
7.1.63. (a) L[ a x2 + b x + c ] = a x2 + (b + 2 a) x + (c + b);

Z x
L−1 [ a x2 + b x + c ] = a x2 + (b − 2 a) x + (c − b + 2 a) = e− x ey p(y) dy.
−∞
 
√1 − √1 √
 2 2
7.2.1. (a)  . (i ) The line y = x; (ii ) the rotated square 0 ≤ x + y, x − y ≤ 2;
√1 √1
2 2 (iii ) the unit disk.
 
3 4
− 5 5
(c)  . (i ) The line 4 x + 3 y = 0; (ii ) the rotated square with vertices
4 3 T √ T
5 5 T
( 0, 0 )T , √1 , √1 , 0, 2 , − √1 , √1 ; (iii ) the unit disk.
2 2 2 2
!
−1 0
7.2.2. (a) L2 = represents a rotation by θ = π;
0 −1
(b) L is clockwise rotation by 90◦ , or, equivalently, counterclockwise rotation by 270◦ .
! !
−1 −4
7.2.5. The image is the line that goes through the image points , .
2 −1

3
2.5
! ! ! !
0 1 4 3 2
7.2.6. (a) Parallelogram with vertices , , , : 1.5
0 2 3 1 1
0.5
1 2 3 4
3
! ! ! !
0 2 3 1
(c) Parallelogram with vertices , , , : 2
0 1 4 3
1
0.5 1 1.5 2 2.5 3
0.75
0.5
0.25
! !
− 12 1 -1 -0.75-0.5-0.25 0.25 0.5 0.75 1
(g) Line segment between 1 and : -0.25
2 −1 -0.5
-0.75
-1
7.2.9. (b) True. (d) False: in general circles are mapped to ellipses.
! ! ! !
1 1 1 0 1 0 1 1
7.2.13. (b) = :
−1 1 −1 1 0 2 0 1
A shear of magnitude −1 along the x-axis, followed by a scaling in the y direction by a fac-
tor of 2, followed by a shear of magnitude −1 along the y-axis.
        
1 1 0 1 0 0 1 0 0 1 0 0 1 0 0 1 0 0 1 1 0
        
(d) 1 0  
1 = 1 1  
00 1 
00 1  
0   0 −1 
00  
1 −1   0 1 0
 :
0 1 1 0 0 1 0 −1 1 0 0 2 0 0 1 0 0 1 0 0 1
A shear of magnitude 1 along the x-axis that fixes the x z-plane, followed a shear of mag-
nitude −1 along the y-axis that fixes the xy plane, followed by a reflection in the xz plane,
followed by a scaling in the z direction by a factor of 2, followed a shear of magnitude −1
along the z-axis that fixes the x z-plane, followed a shear of magnitude 1 along the y-axis
that fixes the yz plane.
 
1 1
7.2.15. (b)  2 2
.
1 1
2 2
 
1 0 0
 
7.2.17. 0 1 0
  is the identity transformation;
0 0 1
 
0 0 1
 
0 1 0
  is a reflection in the plane x = z;
1 0 0
 
0 1 0
  ◦
0 0 1
  is rotation by 120 around the line x = y = z.
1 0 0

 
! −1 0 0
−1 0  ◦ 
7.2.19. det = +1, representing a 180 rotation, while 0 −1 0 = −1, det 

0 −1
0 0 −1
and so is a reflection — but through the origin, not a plane, since it doesn’t fix any nonzero
vectors.
♦ 7.2.21. (a) First, w = (u · v) u = (uT v) u = u uT v is the orthogonal projection of v onto the

line in the direction of u. So the reflected vector is v − 2 w = ( I − 2 u uT )v.
 
7 24
 25 0 − 25 
 
(c) (i )  0
 0 1 .
 
24 7
− 25 0 − 25
! !
1 −6 −1 0
7.2.24. (b) , (d) .
− 34 3 0 5
 
−3 −1 −2
 
7.2.25. (a)  6 1 6
 .
1 1 0
     
1 −4 0 ! ! !
  
0,
   1 2 1 0 0
7.2.26. (b) bases:   

0,4

, and , ; canonical form: ;
−2 1 0 0 0
0 1 3
             
1 0 1 1 2 −1 1 0 0
             
(d) bases:  0  ,  1  ,  −2 , and  1  ,  −1  ,  −1 ; canonical form: 0 1 0
             .
0 0 3 2 1 1 0 0 0
♦ 7.2.28. (a) Let Q have columns u1 , . . . , un , so Q is an orthogonal matrix. Then the matrix rep-
resentative in the orthonormal basis is
B = Q−1 A Q = QT A Q, and B T = QT AT (QT )T = QT A Q = B.
! ! !
1 0 1 −1 2 0
(b) Not necessarily. For example, if A = ,S = , then S −1 A S =
0 2 1 0 1 1
is not symmetric.
7.3.1. (b) True. (d) False: in general circles are mapped to ellipses.
7.3.3. (a) (i ) The horizontal line y = −1; (ii ) the disk (x − 2)2 + (y + 1)2 ≤ 1 of radius 1
centered at ( 2, −1 )T ; (iii ) the square { 2 ≤ x ≤ 3, −1 ≤ y ≤ 0 }.
(c) (i ) The horizontal line y = 2; (ii ) the elliptical domain x2 − 4 x y + 5 y 2 + 6 x − 16 y + 12 ≤ 0;
(iii ) the parallelogram with vertices ( 1, 2 )T , ( 2, 2 )T , ( 4, 3 )T , ( 3, 3 )T .
! !
−2 1 2
7.3.4. (a) T3 ◦ T4 [ x ] = x+ ,
−1 0 2
! ! ! ! ! ! !
−2 1 1 2 0 1 2 1 2 1 1
with = , = + ;
−1 0 0 1 −1 0 2 0 1 0 2

   
3 3
2 2 2
(c) T3 6 [ x ] =
◦T  x +  ,
1 1
2 2 2
     ! ! ! !
3 3 1 1
2 2 1 2 2 2 2 1 2 1 1
with  =  , = + .
1 1
0 1 1 1 2 0 1 0 2
2 2 2 2
7.3.7. (a) F [ x ] = A x + b has an inverse if and only if A in nonsingular.

(b) Yes: F −1 [ x ] = A−1 x − A−1 b.
! ! ! ! ! ! ! !
−1 x 1 −2 x 3 x 0 −1 x 0
(c) T3 = + , T4−1 = + ,
y 0 1 y −2 y 1 0 y −1
T6 has no inverse.
7.3.11. (a) Isometry, (c) not an isometry.

√      √   √ 
3
− 21 1 1 3
− 12 3− 3
7.3.15. (b) F [ x ] =  2 √ x −   +   =  2 √ x +  2√ 
;
1 3 1 3 1− 3
2 2 " 1 1 2 2 2
! !# ! ! !
0 −1 −2 −2 0 −1 −1
G[ x ] = x− + = x+ ;
1 0 1 1 1 0 3
 √   √   √   √   √ 
1 3
− − − 3 − 12 − 23 −1− 3 −1− 3
F ◦ G[ x ] =  √2 2 
x+ √  =  √3 x −  2 √ + 2√ 
3
2 − 12 3 2 − 1
2
−1+ 3
2
−1+ 3
2
 √ 
−1− 3
2√ 
is counterclockwise rotation around the point  by 120◦ ;
−1+ 3
2
 √   √   √   √   √ 
1 3 −3+ 3
− − 1
− 2 − 23 −1− 3 −1− 3
G ◦ F[x] =  √2 2 
x+ 2√ = √ x −  2√  
+ 2√ 
3
2 − 12 9− 3
2 2
3
− 1
2
5− 3
2
5− 3
2
 √ 
−1− 3
2√ 
is counterclockwise rotation around the point  by 120◦ .
5− 3
2
♥ 7.3.16. (a) If F [ x ] = Q x + a and G[ x ] = R x + b, then G ◦ F [ x ] = R Q x + (R a + b) = S x + c

is a isometry since S = Q R, the product of two orthogonal matrices, is also an orthogonal
matrix. (b) F [ x ] = x + a and G[ x ] = x + b, then G ◦ F [ x ] = x + (a + b) = x + c.
   
! ! ! ! √3 √3
1 0 2 x+2 0 1  2 
y+
2
♦ 7.3.17. (a) x+ = , (b) x+   = .
0 −1 0 −y 1 0 √3 x+ √3
2 2
♦ 7.3.21. First, if L is an isometry, and k u k = 1 then k L[ u ] k = 1, proving that L[ u ] ∈ S1 .

Conversely, if L preserves the unit sphere, and 0 6= v ∈ V , then u = v/r ∈ S1 where
r = k v k, so k L[ v ] k = k L[ r v ] k = k r L[ u ] k = r k L[ u ] k = r = k v k, proving (7.40).
2 2
♥ 7.3.24. (a) q(H x) = (x cosh α + y sinh α) − (x sinh α + y cosh α)
= (cosh2 α − sinh2 α)(x2 − y 2 ) = x2 − y 2 = q(x).

7.4.2. (a) L(x) = 3 x; domain R; codomain R; right-hand side −5; inhomogeneous.

! !
u − 2v −3
(c) L(u, v, w) = ; domain R 3 ; codomain R 2 ; right-hand side ;
v−w −1
inhomogeneous.
(e) L[ u ] = u′ (x) + 3 x u(x); domain C1 (R); codomain C0 (R); right-hand side 0;
homogeneous.
!
u′ (x) − u(x)
(g) L[ u ] = ; domain C1 (R); codomain C0 (R) × R;
u(0)
!
0
right-hand side ; inhomogeneous.
1
 ′′ 
u (x) − v ′′ (x) − 2 u(x) + v(x)
 
(k) L[ u, v ] =  u(0) − v(0) ; domain C2 (R) × C2 (R);
 
u(1) − v(1)
 
0
 
codomain C0 (R) × R 2 ; right-hand side  
 0 ; homogeneous.
0
Z ∞
(m) L[ u ] = u(t) e−s t dt; domain C0 (R); codomain C0 (R);
0
right-hand side 1 + s2 ; inhomogeneous.
Z b
7.4.3. L[ u ] = u(x) + K(x, y) u(y) dy. The domain is C0 (R) and the codomain is R. To
a
show linearity, for constants c, d,
Z b
L[ c u + d v ] = [ c u(x) + d v(x) ] + K(x, y) [ c u(y) + d v(y) ] dy
a
Z b ! Z b !
=c u(x) + K(x, y) u(y) dy +d v(x) + K(x, y) v(y) dy = c L[ u ] + d L[ v ].
a a
7.4.6. (a) u(x) = c1 e2 x + c2 e− 2 x , dim = 2; (c) u(x) = c1 + c2 e3 x + c3 e− 3 x , dim = 3.
7.4.9. (a) p(D) = D3 + 5 D2 + 3 D − 9. (b) ex , e− 3 x , x e− 3 x .

The general solution is y(x) = c1 e + c2 e− 3 x + c2 x e− 3 x .
x
7.4.10. (a) Minimal order 2: u′′ + u′ − 6 u = 0. (c) minimal order 2: u′′ − 2 u′ + u = 0.

√ √
c2
7.4.11. (a) u = c1 x + , (c) u = c1 | x |(1+ 5)/2
+ c2 | x |(1− 5)/2
.
x5
♦ 7.4.14. (b) (i ) u(x) = c1 x + c2 x log | x |
e2 x e− 2 x
7.4.15. v ′′ −4 v = 0, so u(x) = c1 +c2 . The solutions with c1 +c2 = 0 are continuously
x x
differentiable at x = 0, but only the zero solution is twice continuously differentiable.
7.4.18. u = c1 + c2 log r. The solutions form a two-dimensional vector space.
7.4.21. (a) Basis: 1, x, y, z, x2 − y 2 , x2 − z 2 , x y, x z, y z; dimension = 9.

   
7
 
0 
− 5 
   
7.4.24. (a) Not in the image. (b) x = 1 +
  z
 − 65 .

   
0 1
7.4.25. (b) x = − 17 + 3
7 z, y = 4
7 + 2
7 z, not unique; (d) u = 2, v = −1, w = 0, unique.
1 x
7.4.26. (b) u(x) = 6e sin x + c1 e2 x/5 cos 45 x + c2 e2 x/5 sin 54 x.
1 1 1 1 11 − x 9 −x
7.4.27. (b) u(x) = 4 − 4 cos 2 x, (d) u(x) = − 10 cos x + 5 sin x + 10 e cos 2 x + 10 e sin 2 x.
7.4.28. (b) u(x) = 1

2 log x + 3
4 + c1 x + c2 x 2 .
√
sin 2 x
7.4.29. (a) Unique solution: u(x) = x − π √ ;
sin 2 π
(d) infinitely many solutions: u(x) = 1
2 + c e− x sin x; (f ) no solution.
3x
7.4.32. (b) u(x) = − 19 x − 1
10 sin x + c1 e + c2 e− 3 x ,
x 1 x 1 −x
(d) u(x) = 1
6 x e − 18 e + 4 e + c1 ex + c2 e− 2 x .
√ √
7.4.35. u(x) = − 7 cos x − 3 sin x.
1 1 3 1
7.4.36. (a) u(x) = 9 x + cos 3 x + 27 sin 3 x, (c) u(x) = 3 cos 2 x + 10 sin 2 x − 5 sin 3 x.
7.4.41. (b) u(x) = c1 e− 3 x cos x + c2 e− 3 x sin x,

√ √ √ √
(d) u(x) = c1 ex/ 2
cos √1 x+c2 e− x/ 2
cos √1 x+c3 ex/ 2
sin √1 x+c4 e− x/ 2
sin √1 x.
2 2 2 2
7.4.42. (a) Minimal order 2: u′′ + 2 u′ + 10 u = 0;

(c) minimal order 5: u(v) + 4 u(iv) + 14 u′′′ + 20 u′′ + 25 u′ = 0.

7.4.43. (b) u(x) = c1 ex + c2 e( i −1) x = c1 ex + c2 e− x cos x + i e− x sin x.
∂u 2 ∂2u 2 2
♥ 7.4.46. (a) = − k2 e− k t+ i k x = ; (c) e− k t
cos k x, e− k t
sin k x.
∂t ∂x2
7.4.48. (a) Conjugated, (d) not conjugated.
♦ 7.4.51. L[ u ] = L[ v ] + i L[ w ] = f , and, since L is real, the real and imaginary parts of this
equation yield L[ v ] = f , L[ w ] = 0.
!  
13
1 −1 − 10
7.5.1. (a) , (c)  7 7 .
2 3 5 15
7 7

!  
3
2 −3 − 52
7.5.2. Domain (a), codomain (b): ; domain (b), codomain (c): 2 ;
4 9 1 10
3 3
 
6
− 17
domain (c), codomain (a): 7 .
5 5
  7 7

1 −2 0
 
7.5.3. (b) 1 0 − 23 
2 
 
2
0 3 2  
 
1 −2 0 
1 −1 −1 
   
7.5.4. Domain (a), codomain (b): 
 1 0 −3 
; domain (b), codomain (c):

 1 0 −1 
.
 
0 2 6 1 4 5
3 3 3
! !
1 0 −1 2 0 −2
7.5.5. Domain (a), codomain (a): ; domain (a), codomain (c): .
3 2 1 8 8 4
♦ 7.5.8. (a) Given u ∈ U, v ∈ V , we have h u , (L + M )∗ [ v ] i = hh (L + M )[ u ] , v ii =

hh L[ u ] , v ii + hh M [ u ] , v ii = h u , L∗ [ v ] i + h u , M ∗ [ v ] i = h u , (L∗ + M ∗ )[ v ] i.
Since this holds for all u ∈ U, v ∈ V , we conclude that (L + M )∗ = L∗ + M ∗ .
(b) h u , (c L)∗ [ v ] i = hh (c L)[ u ] , v ii = c hh L[ u ] , v ii = c h u , L∗ [ v ] i = h u , c L∗ [ v ] i.
7.5.11. In all cases, L = L∗ if and only if its matrix representative A, with respect to the stan-
! !
−1 0 3 0
dard basis, is symmetric. (a) A = = AT , (c) A = = AT .
0 −1 0 3
 
0 1 1
 
7.5.14. (a) a12 = 21 a21 , a13 = 13 a31 , 21 a23 = 13 a32 , (b) Example: 2 1 2 .
3 3 2
7.5.18. (a) (L + L∗ )∗ = L∗ + (L∗ )∗ = L∗ + L. (b) Since L ◦ L∗ = (L∗ )∗ ◦ L∗ , this follows from

Theorem 7.60. (Or it can be proved directly.)
Z b Z b Z b
7.5.21. (a) h Ma [ u ] , v i = Ma [ u(x) ] v(x) dx = a(x) u(x) v(x) dx = u(x) Ma [ v(x) ] dx
a a a
= h u , Ma [ v ] i, proving self-adjointness.
T
1 1
7.5.24. Minimizer: 5, −5 ; minimum value: − 51 .
T
2 1
7.5.26. Minimizer: 3, 3 ; minimum value: −2.
T
7 2 7
7.5.28. (a) Minimizer: 13 , 13 ; minimum value: − 26 .

12 5 T 43
(c) Minimizer: 13 , 26 ; minimum value: − 52 .
1 6
7.5.29. (a) 3, (b) 11 .

Chapter 8: Eigenvalues and Singular Values
8.1.1. (a) u(t) = − 3 e5 t , (b) u(t) = 3 e2(t−1) .
8.1.2. γ = log 2/100 ≈ .0069. After 10 years: 93.3033 gram; after 100 years: 50 gram;
after 1000 years: .0977 gram.
8.1.5. The solution is u(t) = u(0) e1.3 t . To double, we need e1.3 t = 2, so t = log 2/1.3 = .5332.
To quadruple takes twice as long, t = 1.0664.
To reach 2 million, the colony needs t = log 106 /1.3 = 10.6273.
b du
♦ 8.1.7. (a) If u(t) ≡ u⋆ = − , then = 0 = a u + b, hence it is a solution.
a dt
dv b
(b) v = u − u⋆ satisfies = a v, so v(t) = c ea t , and u(t) = c ea t − .
dt a
8.1.8. (a) u(t) = 1

2 + 1
2 e2 t .
h i
8.1.11. (a) u(t) = 1
3 e2 t/7 . (b) One unit: t = log 1/(1/3 − .3333) /(2/7) = 36.0813;
h i
1000 units: t = log 1000/(1/3 − .3333) /(2/7) = 60.2585;
1 2 t/7
(c) One unit: t ≈ 30.2328 solves 3e − .3333 e.2857 t = 1.
1000 units: t ≈ 52.7548 solves 13 e2 t/7 − .3333 e.2857 t = 1000.
Note: The solutions to these nonlinear equations are found by a numerical equation solver,
e.g., the bisection method, or Newton’s method, [ 8 ].
! !
−1 1
8.2.1. (a) Eigenvalues: 3, −1; eigenvectors: , .
1 1
     
1 −1 1
     
(e) Eigenvalues: 4, 3, 1; eigenvectors:  −1 ,  0 ,  2 
    
.
1 1 1
     
3 3− 2i 3+ 2i
     
(g) Eigenvalues: 0, 1 + i , 1 − i ; eigenvectors:  1 ,  3− i ,  3+ i .
     
0 1 1
 
2
 
(i) −1 is a simple eigenvalue, with eigenvector 
 −1 ;

1
1  2
−
3  3
2 is a double eigenvalue, with eigenvectors  0 ,  1
   .
1 0
8.2.4. (a) O is a trivial example.
58
! !
1 −1
8.2.7. (a) Eigenvalues: i , −1 + i ; eigenvectors: , .
0 1
! 3 1 !
−1 5 + 5i
(c) Eigenvalues: −3, 2 i ; eigenvectors: , .
1 1
! !
−1 1
8.2.9. For n = 2, the eigenvalues are 0, 2, and the eigenvectors are , and .
1 1
     
−1 −1 1
     
For n = 3, the eigenvalues are 0, 0, 3, and the eigenvectors are 
 0 ,  1 , and
  
 1 .
 
1 0 1
♦ 8.2.10. (a) If A v = λ v, then A(c v) = c A v = c λ v = λ (c v) and so c v satisfies the eigenvector

equation for the eigenvalue λ. Moreover, since v 6= 0, also c v 6= 0 for c 6= 0, and so c v is a
bona fide eigenvector. (b) If A v = λ v, A w = λ w, then
A(c v + d w) = c A v + d A w = c λ v + d λ w = λ (c v + d w).
8.2.12. True — by the same computation as in Exercise 8.2.10(a), c v is an eigenvector for the
same (real) eigenvalue λ.
8.2.15. (a) tr A = 2 = 3 + (−1); det A = −3 = 3 · (−1).

(e) tr A = 8 = 4 + 3 + 1; det A = 12 = 4 · 3 · 1.
√ √
(g) tr A = 2 = 0 + (1 + i ) + (1 − i ); det A = 0 = 0 · (1 + i 2) · (1 − i 2).
(i) tr A = 3 = (−1) + 2 + 2; det A = −4 = (−1) · 2 · 2.
8.2.17. If U is upper triangular, so is U − λ I , and hence p(λ) = det(U − λ I ) is the product of

Y
the diagonal entries, so p(λ) = (uii − λ). Thus, the roots of the characteristic equation
i
are u11 , . . . , unn — the diagonal entries of U .
! !
0 1 0 0
8.2.21. (a) False. For example, 0 is an eigenvalue of both and , but the eigen-
! 0 0 1 0
0 1
values of A + B = are ± i .
1 0
♦ 8.2.24. (a) Starting with A v = λ v, multiply both sides by A−1 and divide by λ to obtain
A−1 v = (1/λ) v. Therefore, v is an eigenvector of A−1 with eigenvalue 1/λ.
(b) If 0 is an eigenvalue, then A is not invertible.
! !
0 −1 1 0
8.2.29. (b) False. For example, has eigenvalues i , − i , whereas has
1 0 0 −1
eigenvalues 1, −1.
     
7 24 3 4
25 − 25
8.2.31. (a) (ii ) Q =  . Eigenvalues −1, 1; eigenvectors  5 ,  5 .
24 7 4
− 25 − 25 5 − 35

h i
−1 −1
♦ 8.2.32. (a) det(B − λ I ) = det(S A S − λ I ) = det S (A − λ I )S
= det S −1 det(A − λ I ) det S = det(A − λ I ).
(b) The eigenvalues are the roots of the common characteristic equation. (c) Not usually.
If w is an eigenvector of B, then v = S w is an eigenvector of A and conversely.
8.2.36. (a) The characteristic equation of a 3 × 3 matrix is a real cubic polynomial, and hence
 
0 1 0 0
 
 −1 0 0 0
has at least one real root. (b) 
 has eigenvalues ± i .
 0 0 0 1
0 0 −1 0
 
0 0 1 √
  1 3
8.2.39. False. For example, 1 0 0
  has eigenvalues 1, − 2 ± 2 i.
0 1 0
8.2.40. (a) According to Exercise 1.2.29, if z = ( 1, 1, . . . , 1 )T , then A z is the vector of row

sums of A, and hence, by the assumption, A z = z, which means that z is an eigenvector
with eigenvalue 1.
♦ 8.2.44. (a) The axis of the rotation is the eigenvector v corresponding to the eigenvalue +1.
Since Q v = v, the rotation fixes the axis, and hence must rotate around it. Choose an
orthonormal basis u1 , u2 , u3 , where u1 is a unit eigenvector in the direction of the axis of
rotation, while u2 + i u3 is a complex eigenvector for the eigenvalue e i θ . In this basis, Q
 
1 0 0
 
has matrix form  
 0 cos θ − sin θ , where θ is the angle of rotation.
0 sin θ cos θ
 
2
 
(b) The axis is the eigenvector  −5  for the eigenvalue 1. The complex eigenvalue is
 
1
√
7 2 30 7
13 +i 13 , and so the angle is θ = cos−1 13 ≈ 1.00219.
! ! !
0 1 1 −1
♥ 8.2.47. (a) M2 = : eigenvalues 1, −1; eigenvectors , .
1 0 1 1
♥ 8.2.52. (a) Follows by direct computation:

! ! ! !
a2 + b c ab + bd a b 1 0 0 0
pA (A) = − (a + d) + (a d − b c) = .
ac + cd b c + d2 c d 0 1 0 0
(b) By part (a), O = A−1 pA (A) = A − (tr A) I + (det A)A−1 , and the formula follows upon
solving for A−1 . (c) tr A = 4, det A = 7 and one checks A2 − 4 A + 7 I = O.
8.2.53. (a) Gershgorin disk: | z − 1 | ≤ 2; eigenvalues: 3, −1; -2 -1 1 2 3 4
-1
-2

1
(c) Gershgorin disks: | z − 2 | ≤ 3, | z | ≤ 1;
√ -1 1 2 3 4 5
eigenvalues: 1 ± i 2; -1
-2
-3
(e) Gershgorin disks: | z + 1 | ≤ 2, | z − 2 | ≤ 3, 2
| z + 4 | ≤ 3; eigenvalues: −2.69805±.806289 i , 2.3961; -8 -6 -4 -2 2 4 6 8
-2
-4
♦ 8.2.55. (i ) Because A and its transpose AT have the same eigenvalues, which must therefore
belong to both DA and DAT .
2
(ii ) (a) Gershgorin disk: | z − 1 | ≤ 2; eigenvalues: 3, −1; -2 -1 1 2 3 4
-1
-2
1
(c) Gershgorin disks: | z − 2 | ≤ 1, | z | ≤ 3;
√ -3 -2 -1 1 2 3
eigenvalues: 1 ± i 2; -1
-2
-3
(e) Gershgorin disks: | z + 1 | ≤ 2, | z − 2 | ≤ 4, | z + 4 | ≤ 2; 2
eigenvalues: −2.69805±.806289, 2.3961; -8 -6 -4 -2 2 4 6 8
-2
-4
!
1 2
8.2.56. (a) False: is a counterexample.
2 5
8.3.1. (a) Complete; dim = 1 with basis ( 1, 1 )T .

(e) Complete; dim = 2 with basis ( 1, 0, 0 )T , ( 0, −1, 1 )T . (g) Not an eigenvalue.

!
2
8.3.2. (a) Eigenvalue: 2; eigenvector: ; not complete.
1
!
1± i
(c) Eigenvalues: 1 ± 2 i ; eigenvectors: ; complete.
2
   
1 1
   

(e) Eigenvalue 3 has eigenspace basis  ,  0 
1  
; not complete.
0 1
! !
−1 1
8.3.3. (a) Eigenvalues: −2, 4; the eigenvectors , form a basis for R 2 .
1 1
! !
i −i
(b) Eigenvalues: 1 − 3 i , 1 + 3 i ; the eigenvectors , , are not real, so the dimen-
1 1
sion is 0.    
1 0
   
(e) The eigenvalue 1 has eigenvector  0 ; the eigenvalue −1 has eigenvector  0 . The
   
0 1
eigenvectors span a two-dimensional subspace of R 3 .
8.3.4. (a) Complex eigenvector basis; (b) complex eigenvector basis;

(e) no eigenvector basis.
8.3.6. (a) True. The standard basis vectors are eigenvectors.
♦ 8.3.11. As in Exercise 8.2.32, if v is an eigenvector of A then S −1 v is an eigenvector of B.

Moreover, if v1 , . . . , vn form a basis, so do S −1 v1 , . . . , S −1 vn ; see Exercise 2.4.21 for details.
! !
3 3 0 0
8.3.13. In all cases, A = S Λ S −1 . (a) S = , Λ= .
1 2 0 −3
! !
− 35 + 1
5 i − 53 − 1
5 i −1 + i 0
(c) S = , Λ= .
1 1 0 −1 − i
   
0 21 1 0 0 0
   
(e) S = 1 −10 6 , Λ = 0 7 0
   .
0 7 3 0 0 −1
   
−4 3 1 0 −2 0 0 0
   
 −3 2 0 1  0 −1 0 0
(h) S =  , Λ =  .

 0 6 0 0

 0 0 1 0
12 0 0 0 0 0 0 2
     
0 −1 0 i −i 0 i 0 0 −i 1
0
      2i 2
1 
8.3.16. (a) 1 0 0 = 1 1 0 0 −i  
0 2 0
    2 .
1
0 0 1 0 0 1 0 0 1 0 0 2
√
8.3.17. (a) Yes: distinct real eigenvalues −3, 2. (c) No: complex eigenvalues 1, − 12 ± 2
5
i.
! !
1 −1 1+ i 0
8.3.18. In all cases, A = S Λ S −1 . (a) S = , Λ= .
1 1 0 −1 + i
! !
1 − 12 − 12 i 4 0
(c) S = , Λ= .
1 1 0 −1

8.3.19. Use the formula A = S Λ S −1 . For part (e) you can choose any other eigenvalues and
eigenvectors you want to fill in S and Λ.
 
  0 0 4
5 4
 
(a)  3 3 , (e) example:  0 1 0.
8 1
3 3 0 0 −2
!
11 −6
8.3.20. (a) .
18 −10
8.3.25. Let A = S Λ S −1 . Then A2 = I if and only if Λ2 = I , and so all its eigenvalues are ±1.
! ! !
3 −2 1 1
Examples: A = , with eigenvalues 1, −1 and eigenvectors , ; or, even
4 −3 1 2
! ! !
0 1 1 −1
simpler, A = , with eigenvalues 1, −1 and eigenvectors , .
1 0 1 1
8.4.1. (a) {0}, the x-axis, the y-axis, the z-axis, the x y-plane, the x z-plane, the y z-plane, R 3 .
8.4.2. (a) {0}, the line x = y, the line x = −y, R 2 ;

(c) {0}, the three lines spanned by each of the eigenvectors ( 1, 0, 1 )T , ( 1, 0, −1 )T , ( 0, 1, 0 )T ,
the three planes spanned by pairs of eigenvectors, R 3 .
8.4.3. (a) Real: {0}, R 2 . Complex: {0}, the two (complex) lines spanned by each of the
complex eigenvectors ( i , 1 )T , ( − i , 1 )T , C 2 .
(d) Real: {0}, the three lines spanned by each of the eigenvectors ( 1, 0, 0 )T , ( 1, 0, −1 )T ,
( 0, 1, −1 )T , the three planes spanned by pairs of eigenvectors, R 3 . Complex: {0}, the three
(complex) lines spanned by each of the real eigenvectors, the three (complex) planes spanned
by pairs of eigenvectors, C 3 .
8.4.6. (a) True; (c) false.
8.4.8. False.
! !
1 −1 1 1
8.5.1. (b) Eigenvalues: 7, 3; eigenvectors: √ , √ .
2 1 2 1
     
4 4
√
5 2 − 3 − √
   5   5 2 
 3     
(d) Eigenvalues: 6, 1, −4; eigenvectors:  √ ,  4 ,  − 3
√

.
5 2   
   5   5 2 

√1 0 √1
2 2
5 1
√
8.5.2. (a) Eigenvalues 2±2 17 ; positive definite. (c) Eigenvalues 0, 1, 3; positive semi-definite.
8.5.5. (a) The characteristic equation p(λ) = λ2 − (a + d)λ + (a d − b c) = 0 has real roots if and
only if its discriminant is non-negative: 0 ≤ (a + d)2 − 4 (a d − b c) = (a − d)2 + 4 b c, which is
the necessary and sufficient condition for real eigenvalues.
(b) If A is symmetric, then b = c and so the discriminant is (a − d)2 + 4 b2 ≥ 0.

♥ 8.5.7. (a) Let A v = λ v. Using the Hermitian dot product,
λ k v k2 = (A v) · v = vT AT v = vT A v = v · (A v) = λ k v k2 ,
and hence λ = λ, which implies that the eigenvalue λ is real.

! !
2− i −2 + i
(c) (ii ) Eigenvalues 4, −2; eigenvectors: , .
1 5
! 1!
5 1 −3 2 .
8.5.9. (a) Generalized eigenvalues: 3, 2; generalized eigenvectors: ,
1 1
1! !
2 , 1
(c) Generalized eigenvalues: 7, 1; generalized eigenvectors: .
1 0
   
! √1 − √2 √1 √2
−3 4  5 5  5 0 
 5 5
8.5.13. (a) =   ,
4 3 √2 √1 0 −5 − √2 √1
5 5 5 5
   
  √1 − √1 √1 √1 √2 √1
1 1 0  6 2 3  3 0 0  6 6 6 
     

(c)  1 2 1 = 
 

√2 0 − √1 

0 1 0

− √ 1
0 √1

.
 6 3  2 2 

0 1 1
√1 √1 √1 0 0 0 √1 − √1 √1
6 2 3 3 3 3
   
! √1 √1
5 −2 
−  5 0 
− √1 √1
2 2   2 2
8.5.14. (b) =    ,
−2 5 √1 √1 0 −10 √1 √1
2 2 2 2
   
4
  √ − 53 − √ 4 √4 3
√ √1
1 0 4  5 2 5 2  6 0 0  5 2 5 2 2
     
 3 4 3 
(d) 
0 1 3=

√
5 2 5
√  0 1
5 2
0 
 − 35 4
5 0.
4 3 1    
√1 √1 4 3 1
0 0 0 −4 − √ − √ √
2 2 5 2 5 2 2
 
57
25 − 24
25
8.5.15. (a)  . (c) None, since eigenvectors are not orthogonal.
24 43
− 25 25
2 2
8.5.17. (b) 7 √1 x+ √2 y + 11
2 − √2 x + √1 y = 7
5 (x + 2 y)2 + 2
5 (− 2 x + y)2 ,
5 5 5 5
2 2 2
1 √1 x + √1 y + √1 z
(d) 2 + − √1 y + √1 z + 2 − √2 x+ √1 y+ √1 z
3 3 3 2 2 6 6 6
= 16 (x + y + z)2 + 12 (− y + z)2 + 31 (− 2 x + y + z)2 .
√ √
8.5.21. Principal stretches = eigenvalues: 4 + 3, 4 − 3, 1;
√ T √ T
principal directions = eigenvectors: 1, −1 + 3, 1 , 1, −1 − 3, 1 , ( −1, 0, 1 )T .
♦ 8.5.23. (a) Let K = Q Λ QT be its spectral factorization. Then xT K x = yT Λ y where x = Q y.

The ellipse yT Λ y = λ1 y12 + λ2 y12 = 1 has its principal axes aligned with the coordinate axes
q
and semi-axes 1/ λi , i = 1, 2. The map x = Q y serves to rotate the coordinate axes to
align with the columns of Q, i.e., the eigenvectors, while leaving the semi-axes unchanged.

1.5
0.5
! !
√ q
2 −1 1
(b) (ii ) -1.5 -1 -0.5 0.5 1 1.5 ellipse: semi-axes 2, 3, principal axes , .
-0.5
1 1
-1
-1.5
8.5.26. Only the identity matrix is both orthogonal and positive definite. Indeed, if K = K T > 0
is orthogonal, then K 2 = I , and so its eigenvalues are all ± 1. Positive definiteness im-
plies that all the eigenvalues must be +1, and hence its diagonal form is Λ = I . But then
K = Q I QT = I also.
√ √
♦ 8.5.27. (a) Set B = Q Λ QT , where Λ is the diagonal matrix with the square roots of the
eigenvalues of A along the diagonal. Uniqueness follows from the fact that the eigenvectors
and eigenvalues are uniquely determined. (Permuting them does not change the final form
√ 
√ √  2 0 0
1 3 + 1 3−1  √ 
of B.) (b) (i ) ; (iii )  0 5 0
√ √  .
2 3−1 3+1 0 0 3
    √ 
2 −3 √2 − √1
 = 5  5  5 0
8.5.29. (b)
1
  √ ,
1 6 √ √2 0 3 5
5 5
    
0 −3 8 0 − 35 45 1 0 0
    
(d) 1 0 0 = 1 0 0  0
   0 5 .
4 3
0 4 6 0 5 5 0 0 10
♥ 8.5.32. (i ) This follows immediately from the spectral factorization. The rows of Λ QT are
λ1 u T T
1 , . . . , λn un , and formula (8.37) follows from the alternative version of matrix multipli-
cation given in Exercise 1.2.34.
!    
1 2 4
−3 4 − 25
(ii ) (a) = 5 5 5
− 5 5 .
4 3 2 4
− 2 1
5 5 5 5
8.5.33. Maximum: 7; minimum: 3.
8.5.35. (a) Maximum: 3; minimum: −2.

√ √
8+ 5 8− 5
(c) Maximum: 2 = 5.11803; minimum: 2 = 2.88197.
√
8.5.37. (a) 5+ 5
2 = max{ 2 x2 − 2 x y + 3 y 2 | x2 + y 2 = 1 },
√
5− 5
2 = min{ 2 x2 − 2 x y + 3 y 2 | x2 + y 2 = 1 };
(c) 12 = max{ 6 x2 − 8 x y + 2 x z + 6 y 2 − 2 y z + 11 z 2 | x2 + y 2 + z 2 = 1 },
2 = min{ 6 x2 − 8 x y + 2 x z + 6 y 2 − 2 y z + 11 z 2 | x2 + y 2 + z 2 = 1 }
8.5.38. (c) 9 = max{ 6 x2 − 8 x y + 2 x z + 6 y 2 − 2 y z + 11 z 2 | x2 + y 2 + z 2 = 1, x − y + 2 z = 0 }.

♦ 8.5.42. According to the discussion preceding the statement of the Theorem 8.42,
n o
λj = max yT Λ y
k y k = 1, y · e1 = · · · = y · ej−1 = 0 .
Moreover, using (8.36), setting x = Q y and using the fact that Q is an orthogonal matrix
and so (Q v) · (Q w) = v · w for any v, w ∈ R n , we have
xT A x = yT Λ y, k x k = k y k, y · e i = x · vi ,
where vi = Q ei is the ith eigenvector of A. Therefore, by the preceding formula,

n o
λj = max xT A x k x k = 1, x · v1 = · · · = x · vj−1 = 0 .
3 2 1
8.5.46. (a) Maximum: 4 ; minimum: 5 . (c) Maximum: 2; minimum: 2 .
   
√1 √1 2 −2
 2 2
8.6.1. (a) U =  , ∆= ;
1 √1
− √ 0 2
2 2
   
√3 √2 2 15
(c) U = 

13 13 
, ∆= ;
− √2 √3 0 −1
13 13
♦ 8.6.4. If A is symmetric, its eigenvalues are real, and hence its Schur Decomposition is
A = Q ∆ QT , where Q is an orthogonal matrix. But AT = (QT QT )T = Q T T QT , and
hence ∆T = ∆ is a symmetric upper triangular matrix, which implies that ∆ = Λ is a
diagonal matrix with the eigenvalues of A along its diagonal.
♥ 8.6.5. (a) If A is real, A† = AT , and so if A = AT then ATA = A2 = A AT .

(b) If A is unitary, then A† A = I = A A† .
(c) Every real orthogonal matrix is unitary, so this follows from part (b).
8.6.6. (b) Two 1 × 1 Jordan blocks; eigenvalues −3, 6; eigenvectors e1 , e2 .

(d) One 3 × 3 Jordan block; eigenvalue 0; eigenvector e1 .
(e) One 1 × 1, 2 × 2, and 1 × 1 Jordan blocks; eigenvalues 4, 3, 2; eigenvectors e1 , e2 , e4 .
! !
1 0
8.6.7. (a) Eigenvalue: 2; Jordan basis: v1 = , v2 = 1 ;
0 3
!
2 1
Jordan canonical form: .
0 2
     
1 0 0
     
(c) Eigenvalue: 1; Jordan basis: v1 =  0
 
 , v =  1  , v =  −1 ;
2   3  
0 0 1
 
1 1 0
 
Jordan canonical form:  0 1 1 .

0 0 1

     
−1 0 0
     
(e) Eigenvalues: −2, 0; Jordan basis: v1 = 0 , v =  −1  , v =  −1 ;
  2   3  
1 0 1
 
−2 1 0
 
Jordan canonical form:  0 −2 0 
 .
0 0 0
       
2 0 0 0 2 1 0 0 2 0 0 0 2 0 0 0
       
0 2 0 0 0 2 0 0 0 2 1 0 0 2 0 0
 ,  ,  ,  ,

0 0 2 0

0 0 2 0

0 0 2 0

0 0 2 1
8.6.8.
0 0 0 2 0 0 0 2 0 0 0 2 0 0 0 2
       
2 1 0 0 2 1 0 0 2 0 0 0 2 1 0 0
       
0 2 0 0 0 2 1 0 0 2 1 0 0 2 1 0
 ,  ,  ,  .

0 0 2 1

0 0 2 0

0 0 2 1

0 0 2 1
0 0 0 2 0 0 0 2 0 0 0 2 0 0 0 2
8.6.11. True. All Jordan chains have length one, and so consist only of eigenvectors.
8.6.16. True. If zj = c wj , then A zj = cA wj = c λwj + c wj−1 = λzj + zj−1 .

!
0 1
♦ 8.6.19. (a) Let A = . Then e2 is an eigenvector of A2 = O, but not an eigenvector of A.
0 0
(b) Suppose A = S J S −1 where J is the Jordan canonical form of A.
Then A2 = S J 2 S −1 . Now, even though J 2 is not necessarily a Jordan matrix, cf. Exercise
8.6.18, since J is upper triangular with the eigenvalues on the diagonal, J 2 is also upper
triangular and its diagonal entries, which are its eigenvalues and the eigenvalues of A2 , are
the squares of the diagonal entries of J.
♦ 8.6.24. First, since Jλ,n is upper triangular, its eigenvalues are its diagonal entries, and hence λ
is the only eigenvalue. Moreover, v = ( v1 , v2 , . . . , vn )T is an eigenvector if and only if
(Jλ,n − λ I )v = ( v2 , . . . , vn , 0 )T = 0. This requires v2 = · · · = vn = 0, and hence v must be
a scalar multiple of e1 .
8.6.26. (a) {0}, the y-axis, R 2 ; (c) {0}, the line spanned the eigenvector ( 1, −2, 3 )T , the
plane spanned by ( 1, −2, 3 ) , ( 0, 1, 0 )T , and R 3 ;
T
(e) {0}, the x-axis, the w-axis,
4
the x z-plane, the y w-plane, R .
q √ √ √ √
8.7.1. (a) 3± 5; (c) 5 2 ; (e) 7, 2.
√  √  q  √ 
! 5
√ √−1+
√−1− √ 5  √  √−2+ 5
√ √ 1
√ 
1 1 10−2 5  10+2 5  3 + 5 0  10−4 5 10−4 5 

8.7.2. (a)
0
=
2 2  2  q √   √
5 ,
√ √ √ √ 0 3− 5 √−2− √ √ 1 √
10−2 5 10+2 5 10+4 5 10+4 5
!  
1 −2 − √1 √
(c) =  10 ( 5 2 ) − √1 √2 ,
−3 6 √3 5 5
10
  √  
!
− √2 √1 4
0  − √35 − √35
3 √1 √3
2 1 0 −1  5 5  7  35 35 
(e) =  √ .
0 −1 1 1 √1 √2 0 2 √2 − √1 √2 √1
5 5 10 10 10 10

√
8.7.4. (a) The eigenvalues of K = ATA are 15 2 ±
221
2 = 14.933, .0667. The square roots of
these eigenvalues give us the singular values of A. i.e., 3.8643, .2588. The condition number
is 3.86433 / .25878 = 14.9330.
(c) The singular values are 3.1624, .0007273, and so the condition number is
3.1624 / .0007273 = 4348.17; the matrix is slightly ill-conditioned.
♠ 8.7.6. In all cases, the large condition number results in an inaccurate solution.
(a) The exact solution is x = 1, y = −1; with three digit rounding, the computed solution is
x = 1.56, y = −1.56. The singular values of the coefficient matrix are 1615.22, .274885, and
the condition number is 5876.
8.7.8. Let A = v ∈ R n be the matrix (column vector) in question. (a) It has one singular
v vT
value: k v k; (b) P = , Σ = k v k — a 1 × 1 matrix, Q = (1); (c) v+ = .
kvk k v k2
8.7.10. Almost true, with but one exception — the zero matrix.
8.7.13. True. If A = P Σ QT is the singular value decomposition of A, then the transposed

equation AT = Q Σ P T gives the singular value decomposition of AT , and so the diagonal
entries of Σ are also the singular values of AT . Note that the square of the singular values
of A are the nonzero eigenvalues of K = AT A, whereas the square of the singular values of
AT are the nonzero eigenvalues of K f = AAT 6= K. Thus, this result implies that the two
Gram matrices have the same non-zero eigenvalues.

!
1 1 √
8.7.16. False. For example, U = has singular values 3 ± 5.
0 2
8.7.21. False. For example, the 2 × 2 diagonal matrix with diagonal entries 2 · 10k and 10−k for
k ≫ 0 has determinant 2 but condition number 2 · 102 k .
8.7.26. (a) .671855, (c) .9755, (e) 1.1066.

T
8.7.29. (a) k A k = 72 . The “unit sphere” for this norm is the rectangle with corners ± 12 , ± 31 .

5 1 T 1 7 T
It is mapped to the parallelogram with corners ± 6,− 6 , ± 6, 6 , with respective
5 7 7
norms 3 and 2, and so k A k = max{ k A v k | k v k = 1 } = 2.
(c) According to Exercise 8.7.28, k A k is the square root of the largest generalized
! !
2 0 5 −4
eigenvalue of the matrix pair K = , AT K A = .
r 0 3 −4 14
√
43+ 553
Thus, k A k = 12 = 2.35436.
 
  0 0 0  
1 2 1 1 3
5 5  
8.7.33. (b)  , (d)  0  140 70 140 
2 1 0 −1 , (f )
3 3 9
.
− 5 5 1 0 0 140 70 140
   
1 2 9
! 7 7  !  7
1 1 1   5  
8.7.34. (b) A = , A+ = 4
7
5
− 14 ,
 x⋆ = A+ =  15 .
 7 
2 −1 1   2  
2 1 11
7 14 7

8.7.38. We list the eigenvalues of the graph Laplacian; the singular values of the incidence
matrix are obtained by taking square roots.
√ √ √ √
7+ 5 5+ 5 7− 5 5− 5
(i ) 4, 3, 1, 0; (iii ) , , , , 0.
2 2 2 2
8.8.1. Assuming ν = 1: (b) Mean = 1.275; variance = 3.995; standard deviation = 1.99875.
(d) Mean = .4; variance = 2.36; standard deviation = 1.53623.
8.8.2. Assuming ν = 1: (b) Mean = .36667; variance = 2.24327; standard deviation = 1.49775.
(d) Mean = 1.19365; variance = 10.2307; standard deviation = 3.19855.
8.8.4. For this to be valid, we need to take ν = 1/m. Then

1 m
X
σxy = (xi − x) (yi − y)
m i=1
   
1 m
X 1 m
X 1 m
X
= xi yi − x yi − y yi  + x y = x y − x y.
m i=1 m i=1 m i=1
8.8.6. Observe that the row vector containing the column sums of A is obtained by left multi-
plication by the row vector e = ( 1, . . . , 1 ) containing all 1s. Thus the columns sums of A
are all zero if and only if e A = 0. But then clearly, e A B = 0.
♣ 8.8.8. (a) The singular values/principal variances are 31.8966, .93037, .02938, .01335 with
       
.08677 −.80181 .46356 −.06779
       
 −.34555   −.44715   −.05729   .13724 
       
       
principal directions 

.77916 ,  .08688 ,  .31273 ,  −.29037 .
      
       
 −.27110   −.05630   −.18453   −.94302 
−.43873 .38267 .80621 −.05448
(b) The fact that there are only 4 nonzero singular values tells us that the data lies on a
four-dimensional subspace. Moreover, the relative smallness of two smaller singular val-
ues indicates that the data can be viewed as a noisy representation of points on a two-
dimensional subspace.

Chapter 9: Iteration
9.1.1. (a) u(1) = 2, u(10) = 1024, u(20) = 1048576; unstable.

(c) u(1) = i , u(10) = −1, u(20) = 1; stable.
9.1.2. (a) u(k+1) = 1.0325 u(k) , u(0) = 100, where u(k) represents the balance after k years.
(b) u(10) = 1.032510 × 100 = 137.69 dollars.
(c) u(k+1) = (1 + .0325/12) u(k) = 1.002708 u(k) , u(0) = 100, where u(k) represents the
balance after k months. u(120) = (1 + .0325/12)120 × 100 = 138.34 dollars.
log | b | − log | a |
9.1.6. | u(k) | = | λ |k | a | > | v (k) | = | µ |k | b | provided k > , where the
log | λ | − log | µ |
inequality relies on the fact that log | λ | > log | µ |.
9.1.10. Let u(k) represent the balance after k years. Then u(k+1) = 1.05 u(k) + 120, with
u(0) = 0. The equilibrium solution is u⋆ = − 120/.05 = − 2400, and so after k years the
balance is u(k) = (1.05k − 1) · 2400. Then
u(10) = $1, 509.35, u(50) = $25, 121.76, u(200) = $4, 149, 979.40.
3k + (−1)k − 3k + (−1)k
9.1.13. (a) u(k) = , v (k) = ;
2 2
√ √ k √ √ k √ √
(k) ( 5 + 2)(3 − 5) + ( 5 − 2)(3 + 5) (k) (3 − 5)k − (3 + 5)k
(c) u = √ , v = √ .
2 5 2 5
√ ! √ !
√ − 2 √ 2
9.1.14. (a) u(k) = c1 (− 1 − 2)k + c2 (− 1 + 2)k ,
1 1
 √   √ 
√ k 5− i 3 √ k 5+ i 3
(b) u(k) = c1 1
2 + 2
3
i  2 + c2 1
2 − 2
3
i  2  .
1 1
√ ! √ !
5
2 cos 13 k π + 23 sin 31 k π 5
2 sin 13 k π − 23 cos 13 k π
= a1 + a2
cos 31 k π sin 31 k π
9.1.16. (a) It suffices to note that the Lucas numbers are the general Fibonacci numbers (9.16)
when a = L(0) = 2, b = L(1) = 1. (b) 2, 1, 3, 4, 7, 11, 18.
!k ! ! !
4 1 −1 −1 3k 0 −2 −1
9.1.18. (b) = ,
−2 1 1 2 0 2k 1 1
   
 k
1 1 2 
1 1 −1   4k 0 0 
1
3 3
1 1
3 
     
1
(d) 
1 2 1 
 = 1 −2 0 
 0 1 0 
 6 − 31 1
6
.

   
2 1 1
1 1 1 0 0 (−1)k − 1
2 0 1
2
70
  
 
u(k) 1 1 −1   2
4k 
(k) ! ! k !  3
u −1 −1 3  (k) 
 



1

9.1.19. (b) = , (d)  v  =  1 −2 0

 .

v (k) 1 2 −2k+1   3 
w(k)
1 1 1 0
√ √ √ √
(5 − 3 5)(2 + 5)k + (5 + 3 5)(2 − 5)k
9.1.22. (a) u(k) = 4
3 − 1
3 (−2)k , (c) u(k) = .
10
♣ 9.1.26. (a) u(k) = u(k−1) + u(k−2) − u(k−8) .

(b) 0, 1, 1, 2, 3, 5, 8, 13, 21, 33, 53, 84, 134, 213, 339, 539, 857, 1363, 2167, . . . .
T
(c) u(k) = u(k) , u(k+1) , . . . , u(k+7) satisfies u(k+1) = A u(k) where the 8 × 8 coefficient
matrix A has 1’s on the superdiagonal, last row ( −1, 0, 0, 0, 0, 0, 1, 1 ) and all other entries 0.
(d) The growth rate is given by largest eigenvalue in magnitude: λ1 = 1.59, with u(k) ∝ 1.59k .
For more details, see [ 44 ].
♠ 9.1.29. (a)
! !
−1 1
E1 : principal axes: , ; semi-axes: 1, 31 ; area: 1
3 π.
1 1
! !
−1 1
E2 : principal axes: ; ; semi-axes: 1, 91 ; area: 1
9 π.
1 1
! !
−1 1 1 1
E3 : principal axes: ; ; semi-axes: 1, 27 ; area: 27 π.
1 1
! !
−1 1 1 1
E4 : principal axes: ; ; semi-axes: 1, 81 ; area: 81 π.
1 1
9.1.32. v(k) = c1 (α λ1 + β)k v1 + · · · + cn (α λn + β)k vn .
9.1.35. According to Theorem 8.32, the eigenvectors of T are real and form an orthogonal basis
of R n with respect to the Euclidean norm. The formula for the coefficients cj thus follows
directly from (4.8).
9.1.37. Separating the equation into its real and imaginary parts, we find
! ! !
x(k+1) µ −ν x(k)
= .
y (k+1) ν µ y (k) !
1
The eigenvalues of the coefficient matrix are µ ± i ν, with eigenvectors and so the
∓i
solution is
! ! !
x(k) x(0) + i y (0) 1 x(0) − i y (0) 1
= (µ + i ν)k + (µ − i ν)k .
y (k) 2 −i 2 i
Therefore z (k) = x(k) + i y (k) = (x(0) + i y (0) ) (µ + i ν)k = λk z (0) .


9.1.41. (a) u(k) = 2k c1 + 1
2 k c2 , v (k) = 1 k
3 2 c2 ;

(c) u(k) = (−1)k c1 − k c2 + 1
2 k(k − 1) c3 , v (k) = (−1)k c2 − (k + 1) c3 , w(k) = (−1)k c3 .
♥ 9.1.43. (a) The system has an equilibrium solution if and only if (T − I )u⋆ = b. In particular,
if 1 is not an eigenvalue of T , every b leads to an equilibrium solution.
(b) Since v(k+1) = T v(k) , the general solution is
u(k) = u⋆ + c1 λk1 v1 + c2 λk2 v2 + · · · + cn λkn vn ,

where v1 , . . . , vn are the linearly independent eigenvectors and λ1 , . . . , λn the correspond-
ing eigenvalues of T .
     
2 1
(k) 3 k −3 −
(c) (i ) u =  −5  − (−3)k  3
;
−1 1 1
       
−1 1 2 2
 3     k 
(iii ) u(k) =  −  − 3 2  + 15
 2   2 (−2)k    
 3  − 5 (−3)  3 .
−1 0 2 3
√ √ √
5+ 33 5− 33 5+ 33
9.2.1. (a) Eigenvalues: 2 ≃ 5.3723,
≃ − .3723; spectral radius:
2 2 ≃ 5.3723.
√
(d) Eigenvalues: 4, − 1 ± 4 i ; spectral radius: 17 ≃ 4.1231.
√
9.2.2. (a) Eigenvalues: 2 ± 3 i ; spectral radius: 13 ≃ 3.6056; not convergent.
4 3 4
(c) Eigenvalues: 5 , 5 , 0; spectral radius: 5; convergent.
√ √
5+ 73 5− 73
9.2.3. (b) Unstable: eigenvalues 12 ≃ 1.12867, 12 ≃ − .29533;
5 1 1
(d) stable: eigenvalues −1, ± i ; (e) unstable: eigenvalues 4, 4, 4.
9.2.6. A solution u(k) → 0 if and only if the initial vector u(0) = c1 v1 + · · · + cj vj is a linear
combination of the eigenvectors (or more generally, Jordan chain vectors) corresponding to
eigenvalues satisfying | λi | < 1 for i = 1, . . . , j.
9.2.10. Since ρ(c A) = | c | ρ(A), then c A is convergent if and only if | c | < 1/ρ(A). So, techni-
cally, there isn’t a largest c.
9.2.14. (a) False: ρ(c A) = | c | ρ(A).

(c) True, since the eigenvalues of A2 are the squares of the eigenvalues of A.

9.2.16. (a) P 2 = limk → ∞ T k 2 = limk → ∞ T 2 k = P . (b) The only eigenvalues of P
are 1 and 0. Moreover, P must be complete, since if v1 , v2 are the first two vectors in a
Jordan chain, then P v1 = λ v1 , P v2 = λ v2 + v1 , with λ = 0 or 1, but P 2 v2 = λ2 v1 +
2 λ v2 6= P v2 = λ v2 +v1 , so there are no Jordan chains except for the ordinary eigenvectors.
Therefore, P = S diag (1, . . . , 1, 0, . . . 0) S −1 for some nonsingular matrix S.

 
! −1
1  
 −2 .
9.2.23. (a) All scalar multiples of ; (c) all scalar multiples of  
1
1
9.2.24. (a) The eigenvalues are 1, 21 , so the fixed points are stable, while all other solutions go
k T
to a unique fixed point at rate 1
2 . When u(0) = ( 1, 0 )T , then u(k) → 3 3
5, 5 .
(c) The eigenvalues are −2, 1, 0, so the fixed points are unstable. Most solutions, specifically
those with a nonzero component in the dominant eigenvector direction, become unbounded.
However, when u(0) = ( 1, 0, 0 )T , then u(k) = ( −1, −2, 1 )T for k ≥ 1, and the solution
stays at a fixed point.
9.2.26. False: T has an eigenvalue of 1, but convergence requires that all eigenvalues be less
than 1 in modulus.
3 8 8
9.2.30. (a) 4, convergent; (c) 7, inconclusive; (e) 7, inconclusive; (f ) .9, convergent.
9.2.31. (a) .671855, convergent; (c) .9755, convergent; (e) 1.1066, inconclusive.
2
9.2.32. (a) 3, convergent; (c) .9755, convergent; (e) .9437, convergent.
9.2.34. Since k c A k = | c | k A k < 1.
♦ 9.2.37. This follows directly from the fact, proved in Proposition 8.62, that the singular values
of a symmetric matrix are just the absolute values of its nonzero eigenvalues.
9.2.41. For instance, any diagonal matrix whose diagonal entries satisfy 0 < | aii | < 1.
! ! !
0 1 1 2 0 −2
9.2.42. (a) False: For instance, if A = ,S = , then B = S −1 A S = ,
0 1 0 1 0 1
and k B k∞ = 2 6= 1 = k A k∞ . (c) True, since A and B have the same eigenvalues.
T
1 5
9.3.1. (b) Not a transition matrix; (d) regular transition matrix: 6, 6 ;
T
1 1 1
(e) not a regular transition matrix; (f ) regular transition matrix: 3, 3, 3 .
9.3.4. 2004: 37,000 city, 23,000 country; 2005: 38,600 city, 21,400 country; 2006: 39,880
city, 20,120 country; 2007: 40,904 city, 19,096 country; 2008: 41,723 city, 18,277 country;
Eventual: 45,000 in the city and 15,000 in the country.
9.3.7. 58.33% of the nights.
9.3.8. When in Atlanta he always goes to Boston; when in Boston he has a 50% probability of
going to either Atlanta or Chicago; when in Chicago he has a 50% probability of going to
either Atlanta or Boston. The transition matrix is regular because
 
.375 .3125 .3125
 
T4 =  .25 .5625 .5 
 has all positive entries.
.375 .125 .1875
On average he visits Atlanta: 33.33%, Boston 44.44%, and Chicago: 22.22% of the time.

9.3.10. Numbering the vertices from top to bottom and left to right, the transition matrix is
 1 1  1 
0 4 4 0 0 0
  9 
 1 1 1 1  2 
 0 0  
 2 4 2 4  9 
   
 1 1 1 1 2 
 0 0   
 2 4 4 2 9 
T = 

.

The probability eigenvector is 
1


and so the bug spends,
1 1

 0 4 0 0 4 0


9


   
 1 1 1 1 2 


0 4 4 2 0 2

9



1
0 0 0 41 0
4
1
9
on average, twice as much time at the edge vertices as at the corner vertices.
 
1 1 1
3 3 3 
 
9.3.14. The limit is 1 1 1 .
3 3 3 
 
1 1 1
3 3 3
T
1 1
9.3.19. All equal probabilities: z = n, . .. , n .
n
X
9.3.22. True. If v = ( v1 , v2 , . . . , vn )T is a probability eigenvector, then vi = 1 and
n
X i=1
tij vj = λ vi for all i = 1, . . . , n. Summing the latter equations over i, we find
j =1 n
X n
X n
X n
X
λ=λ vi = tij vj = vj = 1,
i=1 i=1 j =1 j =1
since the column sums of a transition matrix are all equal to 1.
n
X
♦ 9.3.24. The ith entry of v is vi = tij uj . Since each tij ≥ 0 and uj ≥ 0, the sum vi ≥ 0
j =1
n
X n
X n
X
also. Moreover, vi = tij uj = uj = 1 because all the column sums of T are
i=1 i,j = 1 j =1
equal to 1, and u is a probability vector.
9.4.1. (a) The eigenvalues are − 12 , 1

3 , so ρ(T ) = 1
2 .
T
(b) The iterates will converge to the fixed point − 16 , 1 at rate 1
2 . Asymptotically, they
come in to the fixed point along the direction of the dominant eigenvector ( −3, 2 )T .
9.4.2. (a) ρ(T ) = 2; the iterates diverge: k u(k) k → ∞ at a rate of 2.
9.4.3. (a) Strictly diagonally dominant; (c) not strictly diagonally dominant;
(e) strictly diagonally dominant.
1
♠ 9.4.4. (a) x = 7 = .142857, y = − 72 = − .285714;
(e) x = −1.9172, y = − .339703, z = −2.24204.
♠ 9.4.5. (c) Jacobi spectral radius = .547723, so Jacobi converges to the solution
x = 87 = 1.142857, y = 197 = 2.71429.

 
! .3333
.7857  
 −1.0000 .
9.4.6. (a) u = , (c) u =  
.3571
1.3333
9.4.8. If A u = 0, then D u = − (L + U ) u, and hence T u = − D−1 (L + U ) u = u, proving that

u is a eigenvector for T with eigenvalue 1. Therefore, ρ(T ) ≥ 1, which implies that T is not
a convergent matrix.
9.4.11. False for elementary row operations of types 1 & 2, but true for those of type 3.
1
♠ 9.4.13. (a) x = 7 = .142857, y = − 72 = − .285714;
(e) x = −1.9172, y = − .339703, z = −2.24204.
9.4.14. (a) ρJ = .2582, ρGS = .0667; Gauss–Seidel converges faster.

(c) ρJ = .5477, ρGS = .3; Gauss–Seidel converges faster.
(e) ρJ = .4541, ρGS = .2887; Gauss–Seidel converges faster.
♠ 9.4.17. The solution is x = .083799, y = .21648, z = 1.21508. The Jacobi spectral radius
is .8166, and so it converges reasonably rapidly to the solution; indeed, after 50 iterations,
x(50) = .0838107, y (50) = .216476, z (50) = 1.21514. On the other hand, the Gauss–
Seidel spectral radius is 1.0994, and it slowly diverges; after 50 iterations, x(50) = −30.5295,
y (50) = 9.07764, z (50) = −90.8959.
♣ 9.4.22. (a) Diagonal dominance requires | z | > 4; (b) The solution is u = (.0115385,
−.0294314, −.0755853, .0536789, .31505, .0541806, −.0767559, −.032107, .0140468, .0115385)T .
It takes 41 Jacobi iterations and 6 Gauss–Seidel iterations to compute the first three deci-
mal places of the solution. (c) Computing the spectral radius, we conclude that the Jacobi
Method converges to the solution whenever | z | > 3.6387, while the Gauss–Seidel Method
converges for z < − 3.6386 or z > 2.
!
1.4
♥ 9.4.24. (a) u = .
.2
(b) The spectral radius is ρJ = .40825 and so it takes about −1/ log10 ρJ ≃ 2.57 iterations
to produce each additional decimal place of accuracy.
(c) The spectral radius is ρGS = .16667 and so it takes about −1/ log 10 ρGS ≃ 1.29 itera-
tions to produce each additional decimal place of accuracy.
   
(n+1) 1−ω − 21 ω 3
ω
(d) u = u(n)
+ . 2
− 13 (1 − ω) ω 61 ω 2 − ω + 1 2
3ω− 2ω
1 2
(e) The SOR spectral radius is minimized when the two eigenvalues of Tω coincide, which
occurs when ω⋆ = 1.04555, at which value ρ⋆ = ω⋆ − 1 = .04555, so the optimal SOR
Method is almost 3.5 times as fast as Jacobi, and about 1.7 times as fast as Gauss–Seidel.
(f ) For Jacobi, about −5/ log10 ρJ ≃ 13 iterations; for Gauss–Seidel, about −5/ log10 ρGS =
7 iterations; for optimal SOR, about −5/ log10 ρSOR ≃ 4 iterations.
(g) To obtain 5 decimal place accuracy, Jacobi requires 12 iterations, Gauss–Seidel requires
6 iterations, while optimal SOR requires 5 iterations.

♣ 9.4.27. (a) x = .5, y = .75, z = .25, w = .5. (b) To obtain 5 decimal place accuracy, Jacobi
requires 14 iterations, Gauss–Seidel requires 8 iterations. One can get very good approxi-
mations of the spectral radii ρJ = .5, ρGS = .25, by taking ratios of entries of successive
iterates, or the ratio of norms of successive error vectors. (c) The optimal SOR Method has
ω = 1.0718, and requires 6 iterations to get 5 decimal place accuracy. The SOR spectral
radius is ρSOR = .0718.
♣ 9.4.31. The Jacobi spectral radius is ρJ = .909657. Using equation (9.76) to fix the SOR para-
meter ω = 1.41307 actually slows down the convergence since ρSOR = .509584 while
ρGS = .32373. Computing the spectral radius directly, the optimal SOR parameter is
ω⋆ = 1.17157 with ρ⋆ = .290435. Thus, optimal SOR is about 13 times as fast as Jacobi,
but only marginally faster than Gauss-Seidel.
(k+1)
♥ 9.4.35. (a) u = u(k) + D−1 r(k) = u(k) − D−1 A u(k) + D−1 b
= u(k) − D−1 (L + D + U )u(k) + D−1 b = − D−1 (L + U )u(k) + D−1 b,
which agrees with (9.55).
♠ 9.5.1. In all cases, we use the normalized version (9.82) starting with u(0) = e1 ; the answers
are correct to 4 decimal places. (a) After 17 iterations, λ = 2.00002, u = ( −.55470, .83205 )T .
(c) After 38 iterations, λ = 3.99996, u = ( .57737, −.57735, .57734 )T .
(e) After 36 iterations, λ = 5.54911, u = ( −.39488, .71005, .58300 )T .
♠ 9.5.2. In each case, to find the dominant singular value of a matrix A, we apply the
Power Method to K = ATA and take the square root of its dominant eigenvalue to
q
find the dominant singular value σ1 = λ1 of A.
!
2 −1
(a) K = ; after 11 iterations, λ1 = 13.0902 and σ1 = 3.6180;
−1 13
 
5 2 2 −1
 
 2 8 2 −4 
(c) K = 

; after 16 iterations, λ = 11.6055 and σ = 3.4067.
 2 2 1 −1  
1 1
−1 −4 −1 2
1
♦ 9.5.5. (a) If A v = λ v then A−1 v = v, and so v is also the eigenvector of A−1 .
λ
(b) If λ1 , . . . , λn are the eigenvalues of A, with | λ1 | > | λ2 | > · · · > | λn | > 0 (recalling that
1 1
0 cannot be an eigenvalue if A is nonsingular), then ,..., are the eigenvalues of A−1 ,
λ1 λn
1 1 1 1
and > > ··· > and so is the dominant eigenvalue of A−1 . Thus,
| λn | | λn−1 | | λ1 | λn
applying the Power Method to A−1 will produce the reciprocal of the smallest (meaning the
one closest to 0) eigenvalue of A and its corresponding eigenvector.
(c) The rate of convergence of the algorithm is the ratio | λn /λn−1 | of the moduli of the
smallest two eigenvalues.

♠ 9.5.6. (a) After 15 iterations, we obtain λ = .99998, u = ( .70711, −.70710 )T .

(c) After 12 iterations, we obtain λ = 1.00001, u = ( .40825, .81650, .40825 )T .
(e) After 7 iterations, we obtain λ = −.88536, u = ( −.88751, −.29939, .35027 )T .
♠ 9.5.8. (a) After 11 iterations, we obtain ν ⋆ = 2.00002, so λ⋆ = 1.0000, u = ( .70711, −.70710 )T .

(c) After 10 iterations, ν ⋆ = 2.00000, so λ⋆ = 1.00000, u = ( .40825, .81650, .40825 )T .
! !
.3310 .9436
9.5.11. (a) Eigenvalues: 6.7016, .2984; eigenvectors: , .
.9436 −.3310
     
.2726 .9454 −.1784
     
(c) Eigenvalues: 4.7577, 1.9009, −1.6586; eigenvectors:      
 .7519 ,  −.0937 ,  .6526 .
.6003 −.3120 −.7364
! !
−2 −1
9.5.13. (a) Eigenvalues: 2, 1; eigenvectors: , .
3 1
(c) Eigenvalues: 3.5842, −2.2899, 1.7057;
     
−.4466 .1953 .7491
     
eigenvectors:  −.7076 ,  −.8380 ,  −.2204 .
     
.5476 −.5094 .6247
9.5.15. (a) It has eigenvalues ± 1, which have the same magnitude. The Q R factorization is
trivial, with Q = A and R = I . Thus, R Q = A, and so nothing happens.
√
(b) It has a pair of complex conjugate eigenvalues of modulus 7 and a real eigenvalue −1.
However, running the Q R iteration produces a block upper triangular matrix with the real
eigenvalue at position (3, 3) and a 2 × 2 upper left block that has the complex conjugate
eigenvalues of A as eigenvalues.
9.5.18. (a)
   
1 0 0 8.0000 7.2801 0
   
H= 0 −.9615 .2747   7.2801 3.5660 
 , T = H AH =  20.0189 .
0 .2747 .9615 0 3.5660 4.9811
♠ 9.5.19. (a) Eigenvalues: 24, 6, 3.
9.5.21. (b)
   
1 0 0 0 3.0000 −2.2361 −1.0000 0
   
0 −.8944 0 −.4472   −2.2361 3.8000 2.2361 .4000 
H1 =  , A1 =  ,

0 0 1 0  

 0 1.7889 2.0000 −5.8138 

0 −.4472 0 .8944 0 1.4000 −4.4721 1.2000
   
1 0 0 0 3.0000 −2.2361 .7875 .6163
   
0 1 0 0   −2.2361 3.8000 −2.0074 −1.0631 
H2 =  , A2 =  .

0 0 −.7875 −.6163 


 0 −2.2716 −3.2961 2.2950 

0 0 −.6163 .7875 0 0 .9534 6.4961

♠ 9.5.22. (b) Eigenvalues: 7., 5.74606, −4.03877, 1.29271.
     
√1 √1 √1
 5   5   5 
     
9.6.1. (b) V (1) :  − √2

,
 V (2) :  − √2

,
 V (3) :  − √2

;

 5   5   5 
0 0 0
           
1 1 0 1 0 0
           
0 0 1 0 1  0
(d) V (1) : 
 ,
 V (2) :    
 ,  , V (3) :     
 ,  , 
.

0 0 0 0 0  1
0 0 0 0 0 t0
9.6.2. (a) Let λ = µ + i ν. Then the eigenvector equation A v = λ v implies A x = µ x − ν y,

A y = ν x + µ y. Iterating, we see that every iterate Ak x and Ak y is a linear combination
of x and y, and hence each V (k) is spanned by x, y and hence two-dimensional, noting that
x, y are linearly independent according to Exercise 8.3.12(a).
9.6.7. In each case, the last uk is the actual solution, with residual rk = f − K uk = 0.
! ! ! !
2 .76923 .07692 .78571
(a) r0 = , u1 = , r1 = , u2 = ;
1 .38462 −.15385 .35714
       
−1 −.13466 2.36658 −.13466
       
(c) r0 =   
 −2 , u1 =  −.26933 ,
 r1 =   
 −4.01995 , u2 =  −.26933 ,

7 .94264 −.81047 .94264

   
.72321 .33333
   
r2 =   
 .38287 , u3 =  −1.00000 .

.21271 1.33333
♣ 9.6.8. Remarkably, after only two iterations, the method finds the exact solution: u3 = u⋆ =
( .0625, .125, .0625, .125, .375, .125, .0625, .125, .0625 )T , and hence the convergence is dramat-
ically faster than the other iterative methods.
k rk k2 xT A2 x − 2 bT A xk + k b k2
♦ 9.6.14. dk = T
= Tk 3 k .
rk A rk xk A xk − 2 b T K 2 xk + b T A b
♠ 9.7.1. (a) The coefficients are c0 = 1

2 and cj,k = − 2−j−2 for all j ≥ 0 and k = 0, . . . , 2j − 1.
(b)
r=2 r=5 r = 10

For any integer 0 ≤ j ≤ 2r+1 − 1, on the interval j 2−r−1 < x < (j + 1)2−r−1 , the value of
the Haar approximant of order r is constant and equal to the value of the function f (x) = x
at the midpoint:

sr (x) = f (2 j + 1) 2−r−2 = (2 j + 1) 2−r−2 .
(Indeed, this is a general property of Haar wavelets.) Hence | hk (x) − f (x) | ≤ 2−r−2 → 0 as
r → ∞, proving convergence.
(c) Maximal deviation: r = 2 : .0625, r = 5 : , .0078125 r = 10 : .0002441.
♠ 9.7.2. (i ) (a) The coefficients are c0 = − 61 , c0,0 = 0, cj,k = (2j − 1 − 2 k) 2−2 j−2 for
k = 0, . . . , 2j − 1 and j ≥ 1.
(b)
r=2 r=5 r = 10
In general, the Haar wavelet approximant sr (x) is constant on each subinterval k 2− r−1 <
x < (k + 1) 2− r−1 for k = 0, . . . , 2r+1 − 1 and equal to the value of the function f (x)
at the midpoint. This implies that the maximal error on each interval is bounded by the
deviation of the function from its value at the midpoint, which suffices to prove convergence
sr (x) → f (x) at r → ∞ provided f (x) is continuous.
(c) Maximal deviation: r = 2 : .05729, r = 5 : , .007731 r = 10 : .0002441.
♥ 9.7.4. (b)
   
1 1 1 0 1 1 1 1
   
1 1 −1 0 1 1 1 −1 −1 
W2 = 

, W2−1 =  .
1 −1 0 1 4 
2 −2 0 0
1 −1 0 −1 0 0 2 −2
9.7.6. The function ϕ(x) = σ(x) − σ(x − 1) is merely a difference of two step functions — a box
function. The mother wavelet is thus

1
w(x) = ϕ(2 x) − ϕ(2 x − 1) = σ(x) − 2 σ x − 2 + σ(x − 1)

♠ 9.7.8.
(a) Exercise 9.7.1: (a) The coefficients are c0 = .5181, c0,0 = −.1388, c1,0 = .05765, c1,1 =
−.1833, c2,0 = c2,1 = 0, c2,2 = .05908, c2,3 = −.2055, c3,0 = c3,1 = c3,2 = c3,3 = c3,4 =
c3,5 = −.01572, c3,6 = .05954, c3,7 = −.2167.
(b)
r=2 r=5 r = 10
In general, on each interval [ k 2− j , k 2− j ] for k = 1, . . . , 02j − 1, the Haar wavelet approx-

imant is constant, equal to the value of the function at the midpoint of the interval, and
hence the error on each interval is bounded by the deviation of the function from the value
at its midpoint. Thus, provided the function is continuous, convergence is guaranteed.
(c) Maximal deviation: r = 2 : .2116, r = 5 : .2100, r = 10 : .0331. Note that the deviation
is only of the indicated size near the end of the interval owing to the jump discontinuity at
x = 1 and is much closer elsewhere.
♦ 9.7.10. For the box function, in view of (9.125),
 

 1, 0 < x ≤ 21 , 
 1, 1
2 < x ≤ 1,
ϕ(2 x) =  , ϕ(2 x − 1) = 
 
0, otherwise, 0, otherwise,
hence
(
1, 0 < x ≤ 1,
ϕ(x) = = ϕ(2 x) + ϕ(2 x − 1),
0, otherwise,
proving (9.139).
♦ 9.7.12. We compute
ψ(x) = ϕ(x + 1) = ϕ(2 (x + 1) − 1) + ϕ(2 (x + 1) − 2) = ϕ(2 x + 1) + ϕ(2 x) = ψ(2 x) + ψ(2 x − 1),
9.7.15. Since the inner product integral is translation invariant, by (9.147),

Z ∞
h ϕ(x − l) , ϕ(x − m) i = ϕ(x − l) ϕ(x − m) dx
−∞
Z ∞
= ϕ(x) ϕ(x + l − m) dx = h ϕ(x) , ϕ(x + l − m) i = 0
−∞
provided l 6= m.
9.7.18. To four decimal places: (a) 1.8659, (b) 1.27434.
9.7.21. Almost true — the column sums of the coefficient matrix are both 1; however, the (2, 1)
entry is negative, which is not an allowed probability in a Markov process.
9.7.25. (a) 2−1 + 2−2 ; (c) 1 + 2−2 + 2−3 + 2−5 + 2−7 .

Chapter 10: Dynamics
!
du 0 1
10.1.1. (i ) (a) u(t) = c1 cos 2 t + c2 sin 2 t. (b) = u.
dt −4 0
!
c1 cos 2 t + c2 sin 2 t
(c) u(t) = .
− 2 c1 sin 2 t + 2 c2 cos 2 t
0.4
0.2
(d) (e) 2 4 6 8
-0.2
-0.4
! !
−2t 2t du 0 1 c1 e− 2 t + c2 e2 t
(ii ) (a) u(t) = c1 e + c2 e . (b) = u. (c) u(t) = .
dt 4 0 − 2 c1 e− 2 t + 2 c2 e2 t
7
6
5
4
(d) (e) 3
2
1
-1 -0.5 0.5 1
dv du
♦ 10.1.4. (a) Use the chain rule to compute =− (− t) = − A u(− t) = − A v.
dt dt
(b) Since v(t) = u(− t) parameterizes the same curve as u(t), but in the reverse direction.
! !
dv 0 −1 c1 cos 2 t − c2 sin 2 t
(c) (i ) = v; solution: v(t) = .
dt 4 0 2 c1 sin 2 t + 2 c2 cos 2 t
! !
dv 0 −1 c1 e2 t + c2 e− 2 t
(ii ) = v; solution: v(t) = .
dt −4 0 − 2 c1 e2 t + 2 c2 e− 2 t
10.1.8. False. If u = A u then the speed along the trajectory at the point u(t) is k A u(t) k.

So the speed is constant only if k A u(t) k is constant. (Later, in Lemma 10.29, this will be
shown to correspond to A being a skew-symmetric matrix.)
♠ 10.1.10. In all cases, the t axis is plotted vertically, and the three-dimensional solution curves
(u(t), u(t), t)T project to the phase plane trajectories (u(t), u(t))T .

(i ) The solution curves are helices going around the t axis:
81
(ii ) Hyperbolic curves going away from the t axis in both directions:
7 −5t 8 5t −5t 4 5t
10.1.11. u(t) = 5e + 5e , v(t) = − 14
5 e + 5e .
10.1.12. (b) x1 (t) = − c1 e− 5 t + 3 c2 e5 t , x2 (t) = 3 c1 e− 5 t + c2 e5 t ;

(d) y1 (t) = − c1 e− t − c2 et − 2
3 c3 , y2 (t) = c1 e− t − c2 et , y3 (t) = c1 e− t + c2 et + c3 .

1 2−2 t 1 − 2+2 t 1 − 2+2 t T
10.1.13. (a) u(t) = 2e + 2e , − 21 e2−2 t + 2e ,
√ √ T
(c) u(t) = et cos 2 t, − √1 et sin 2 t ,
2
(e) u(t) = ( − 4 − 6 cos t − 9 sin t, 2 + 3 cos t + 6 sin t, − 1 − 3 sin t )T ,

√ √ √ √ √
10.1.15. x(t) = et/2 cos 2
3
t− 3 sin 2
3
t , y(t) = et/2 cos 2
3
t− √1 sin 2
3
t ,
3
and so at time t = 1, the position is ( x(1), y(1) )T = ( −1.10719, .343028 )T .
10.1.19. The general complex solution to the system is

     
−1 1 1
  (1+2 i ) t   (1−2i) t  
u(t) = c1 e−t 


1  + c2 e  
 i  + c3 e
 
−i .
1 1 1
Substituting into the initial conditions,  
 
2 c1 = − 2,
− c1 + c2 + c3 
   
u(0) =    
 c1 + i c2 − i c3  =  −1  we find c2 = − 21 i ,
 
c1 + +c2 + c3 c3 = 1
i.
−2 2
Thus, we obtain the same solution:
       
−1 1 1 2 e−t + et sin 2 t
       
u(t) = − 2 e−t 
 1−
1
2 i e(1+2 i ) t 
 i+
1
2 i e(1−2i) t  
−i  =  −2 e

−t t
+ e cos 2 t 
.
−t t
1 1 1 −2 e + e sin 2 t
10.1.20. (a) Linearly independent; (b) linearly independent; (d) linearly dependent;
(e) linearly independent.
dv du
10.1.24. =S = S A u = S A S −1 v = B v.
dt dt
♦ 10.1.25. (i ) This is an immediate consequence of the preceding two exercises.
! ! √ √ ! √ 
−1 1 c1 e− 2 t − 2i 2 i  c1 e(1+ i √2) t 
(ii ) (a) u(t) = , (c) u(t) = ,
1 1 c2 e2 t 1 1 c2 e(1− i 2)t
  
4 3 + 2i 3− 2i c1
  
(e) u(t) =  −2
 −2 − i −2 + i 
 c2 e i t 
.
1 1 1 c3 e− i t

! !
c1 e2 t + c2 t e2 t c1 e− 3 t + c2 21 + t e− 3 t
10.1.26. (a) , (c) ,
c2 e2 t 2 c1 e− 3 t + 2 c2 t e− 3 t
 
c1 e− 3 t + c2 t e− 3 t + c3 1 + 21 t2 e− 3 t
 

(e)  c2 e− 3 t + c3 t e− 3 t .

−3t −3t 1 2 −3t
c1 e + c2 te + 2 c3 t e
 
! ! 2 0 0
du 2 − 21 du 0 0 du  
10.1.27. (a) = u, (c) = u, (e) =
0 −3 0 u.
dt 0 1 dt 1 0 dt
2 3 0
dui
10.1.28. (a) No, since neither , i = 1, 2, is a linear combination of u1 , u2 . Or note that the
dt
trajectories described by the functions cross, violating uniqueness.
(b) No, since polynomial solutions a two-dimensional system can be at most first order in t.
! !
−1 0 2 3
(d) Yes: u = u. (e) Yes: u =

u.
0 1 −3 2
♦ 10.1.32. (a) By direct computation,

j j−1
duj X tj−i X tj−i−1
= λ eλ t w + eλ t w,
dt i=1 (j − i) ! i i=1 (j − i − 1) ! i
which equals
 
j j
λt X tj−i tj−1 X tj−i
A uj = e A wi = eλ t  w1 + (λ wi + wi−1 )  .
i=1 (j − i) ! (j − 1) ! i = 2 (j − i) !
(b) At t = 0, we have uj (0) = wj , and the Jordan chain vectors are linearly independent.
10.2.1. (a) Asymptotically stable: the eigenvalues are −2 ± i ;

(d) stable: the eigenvalues are ±4 i ; (f ) unstable: the eigenvalues are 1, −1 ± 2 i .
10.2.3. (a) u = − 2 u, v = − 2 v, with solution u(t) = c1 e− 2 t , v(t) = c2 e− 2 t .

(c) u = − 8 u + 2 v, v = 2 u − 2 v, with solution

√ √ √ √ √ √
13+3 − (5+ 13)t
u(t) = − c1 2 e +c2 13−3
2 e− (5− 13)t
, v(t) = c1 e− (5+ 13)t
+c2 e− (5− 13)t
.
10.2.5. (a) Gradient flow; asymptotically stable. (b) Neither; unstable.

(d) Hamiltonian flow; stable.
10.2.9. The system is stable since ± i must be simple eigenvalues. Indeed, any 5 × 5 matrix
has 5 eigenvalues, counting multiplicities, and the multiplicities of complex conjugate eigen-
values are the same. A 6 × 6 matrix can have ± i as complex conjugate, incomplete double
eigenvalues, in addition to the simple real eigenvalues −1, −2, and in such a situation the
origin would be unstable.
10.2.14. (a) True, since the sum of the eigenvalues equals the trace, so at least one must be
positive or have positive real part in order that the trace be positive.
10.2.20. False. Only positive definite Hamiltonian functions lead to stable gradient flows.

♥ 10.2.22. (a) By the multivariable calculus chain rule

!
d ∂H du ∂H dv ∂H ∂H ∂H ∂H
H(u(t), v(t)) = + = + − ≡ 0.
dt ∂u dt ∂v dt ∂u ∂v ∂v ∂u
Therefore H(u(t), v(t)) ≡ c is constant, with its value
c = H(u0 , v0 ) fixed by the initial conditions u(t0 ) = u0 , v(t0 ) = v0 . 1
(b) The solutions are 0.75
0.5
u(t) = c1 cos(2 t) − c1 sin(2 t) + 2 c2 sin(2 t), 0.25
v(t) = c2 cos(2 t) − c1 sin(2 t) + c2 sin(2 t), -1 -0.75-0.5-0.25 0.25 0.5 0.75 1
-0.25
and leave the Hamiltonian function constant: -0.5
H(u(t), v(t)) = u(t)2 − 2 u(t) v(t) + 2 v(t)2 = c21 − 2 c1 c2 + 2 c22 = c.

-0.75
-1
⋆ ⋆
10.2.24. (a) The equilibrium solution satisfies A u = − b, and so v(t) = u(t) − u satisfies
⋆
v = u = A u + b = A(u − u ) = A v,

which is the homogeneous system.

(b) (i ) u(t) = − 3 c1 e2 t + c2 e− 2 t − 14 , v(t) = c1 e2 t + c2 e− 2 t + 14 .
!
−2 3
10.3.1. (ii ) A = ;
−1 1 √ !
√ 3 3
1 3 2 +i 2
λ1 = 2 +i 2 , v1 = ,
√ !
1
√ 3 3
1 3 −i
λ2 = 2− i 2 , v2 = 2 2 ,
1
√ √ √ √
u1 (t) = e− t/2 3 3
2 c1 − 2 c2 cos 2 t +
3 3
2 c1 + 3
2 c2 sin 2
3
t ,
√ √
u2 (t) = e− t/2 c1 cos 23 t + c2 sin 23 t ;
stable focus; asymptotically stable
! !
2 cos t − sin t 2 sin t + cos t
10.3.2. (ii ) u(t) = c1 e− t + c2 e− t ;
5 cos t 5 sin t
stable focus; asymptotically stable.
!
−1 4
10.3.3. (a) For the matrix A = ,
1 −2
tr A = −3 < 0, det A = −2 < 0, ∆ = 17 > 0,
so this is an unstable saddle point.

!
5 4
(c) For the matrix A = ,
1 2
tr A = 7 > 0, det A = 6 > 0, ∆ = 25 > 0,
so this is an unstable node.
10.3.5. (b) (d)
  !
4 t 1 −2t
e − 3e − 13 et + 1 −2t
3e cos t − sin t
10.4.1. (a) 3 , (c) ,
4 t 4 −2t
− 13 et + 4 −2t sin t cos t
3e − 3e 3 e
2t !
e cos t − 3 e2 t sin t 2t
2 e sin t
(e) .
−5 e2 t sin t e cos t + 3 e2 t sin t
2t
 
 
1 0 0 
e− 2 t + t e− 2 t t e− 2 t t e− 2 t 
   
10.4.2. (a)  2 sin t cos t sin t   −2t −2t − 2 t .
 , (c)  −1+e e −1 + e 
 
2 cos t − 2 − sin t cos t −2t −2t
1−e − te −t e− 2 t 1 − te−2t
10.4.3. Exercise 10.4.1 (a) det et A = e− t = et tr A , (c) det et A = 1 = et tr A , (e) det et A =

e4 t = et tr A . Exercise 10.4.2 (a) det et A = 1 = et tr A , (c) det et A = e− 4 t = et tr A .
! ! !
3 e− t − 2 e− 3 t −3 e− t + 3 e− 3 t −1 −6 e− t + 5 e− 3 t
10.4.4. (b) u(t) = = .
2 e− t − 2 e− 3 t −2 e− t + 3 e− 3 t 1 −4 e− t + 5 e− 3 t
  
√ √ √  e 0 0
 e cos 2 − 2 e sin 2   
10.4.5. (b)  √ √ , (d) 0
 e−2 0 .

√1 e sin 2 e cos 2
2 0 0 e−5
10.4.8. There are none, since et A is always invertible.
10.4.11. The origin is an asymptotically stable if and only if all solutions tend to zero as t → ∞.
Thus, all columns of et A tend to 0 as t → ∞, and hence limt → ∞ et A = O. Conversely, if
limt → ∞ et A = O, then any solution has the form u(t) = et A c, and hence u(t) → 0 as
t → ∞, proving asymptotic stability.
10.4.13. (a) False, unless A−1 = − A.

10.4.15. Set U (t) = A et A , V (t) = et A A. Then, by the matrix Leibniz formula (10.41),
U = A2 et A = A U, V = A et A A = A V,

while U (0) = A = V (0). Thus U (t) and V (t) solve the same initial value problem, hence, by
uniqueness, U (t) = V (t) for all t. Alternatively, one can use the power series formula
(10.47): ∞ tn
X
A et A = An+1 = et A A.
n=0 n!
10.4.20. Lemma 10.28 implies det et A = et tr A = 1 for all t if and only if tr A = 0. (Even if tr A
is allowed to be complex, by continuity the only way this could hold for all t is if tr A = 0.)
d
♦ 10.4.25. (a) diag (et d1 , . . . et dn ) = diag (d1 et d1 , . . . , dn et dn ) = D diag (et d1 , . . . , et dn ).
dt
Moreover, at t = 0, we have diag (e0 d1 , . . . , e0 dn ) = I . Therefore, diag (et d1 , . . . , et dn )
satisfies the defining properties of et D .
! ! !−1  
4 t 1 −2t
1 1 et 0 1 1 e − 3e − 31 et + 1 −2t
3e
(c) 10.4.1: (a) = 3 ;
1 4 0 e− 2 t 1 4 4 t 4 −2t
− 31 et + 4 −2t
3e − 3e 3e
! ! !−1 !
i −i eit 0 i −i cos t − sin t
(c) = ;
1 1 0 e− i t 1 1 sin t cos t
! ! !−1
3
5 − 51 i 3
5 + 51 i e(2+ i )t 0 3
5 − 51 i 3
5 + 51 i
(e)
1 1 0 e(2− i )t 1 1
!
e2 t cos t − 3 e2 t sin t 2 e2 t sin t
= .
−5 e2 t sin t e cos t + 3 e2 t sin t
2t
10.4.2: (a)
   −1  
−1 0 0 1 0 0 −1 0 0 1 0 0
     

 0 −i i 
0 eit 0 
 0 −i i = 
 2 sin t cos t sin t 
;
2 1 1 0 0 e− i t 2 1 1 2 cos t − 2 − sin t cos t
(c) not diagonalizable.
dU
♦ 10.4.28. (a) If U (t) = C et B , then = C et B B = U B, and so U satisfies the differential
dt
equation. Moreover, C = U (0). Thus, U (t) is the unique solution to the initial value prob-

lem U = U B, U (0) = C, where the initial value C is arbitrary.
!
1 0
10.4.32. (b) — shear transformations in the y direction. The trajectories are lines
t 1
parallel to the y-axis. Points on the y-axis are fixed.
!
cos 2 t − sin 2 t
(d) — elliptical rotations around the origin. The trajectories are the
2 sin 2 t 2 cos 2 t
ellipses x2 + 1
4 y 2 = c. The origin is fixed.

 
1
0 t
 
10.4.33. (b) 01 0
  — shear transformations in the x direction, with magnitude propor-
0 0 1
tional to the z coordinate. The trajectories are lines parallel to the x axis. Points on the xy
plane are fixed.
 
cos t sin t 0
 
(d)  − sin t cos t 0 
  — spiral motions around the z axis. The trajectories are the pos-
t
0 0 e
itive and negative z axes, circles in the xy plane, and cylindrical spirals (helices) winding
around the z axis while going away from the xy pane at an exponentially increasing rate.
The only fixed point is the origin.
10.4.35.
" ! !# ! " ! !# !
2 0 0 0 0 0 2 0 0 3 0 6
, = , , = ,
0 0 1 0 −2 0 0 0 −3 0 6 0
" ! !# ! " ! !# !
2 0 0 −1 0 −2 2 0 0 1 0 2
, = , , = ,
0 0 4 0 −8 0 0 0 1 0 −2 0
dU dunj
10.4.41. In the matrix system = A U , the equations in the last row are = 0 for
dt dt
j = 1, . . . , n, and hence the last row of U (t) is constant. In particular, for the exponen-
tial matrix solution U (t) = et A the last row must equal the last row of the identity matrix
U (0) = I , which is eT
n.
!
x+t
♦ 10.4.43. (a) : translations in x direction.
y
! !
(x + 1) cos t − y sin t − 1 −1
(c) : rotations around the point .
(x + 1) sin t + y cos t 0
10.4.44. (a) If U 6= {0}, then the system has eigenvalues with positive real part corresponding
to exponentially growing solutions and so the origin is unstable. If C 6= {0}, then the sys-
tem has eigenvalues with zero real part corresponding to either bounded solutions, which
are stable but not asymptotically stable modes, or, if the eigenvalue is incomplete, polyno-
mial growing unstable modes.
10.4.45. (a) S = U = ∅, C = R 2 ;
√ √
2− 7 ! 2+ 7 !
(b) S = span 3
U = span , C = ∅; 3 ,
1 1
     
1 1 2
     
(d) S = span      
 −1 , U = span  1 , C = span  0 .
−1 −1 −3
10.4.47. (b) u1 (t) = et−1 − et + t et , u2 (t) = et−1 − et + t et ;

13 4 t 3 1 13 4 t 29 3
(d) u(t) = 16 e + 16 − 4 t, v(t) = 16 e − 16 + 4 t.

10.4.48. (a) u1 (t) = 1 1 1 1

2 cos 2 t + 4 sin 2 t + 2 − 2 t,
u2 (t) = 2 e− t − 1
2 cos 2 t − 1
4 sin 2 t − 3
2 + 3
2 t,
−t 1 3 7 3
u3 (t) = 2 e − 4 cos 2 t − 4 sin 2 t − 4 + 2 t.
q
10.5.1. The vibrational frequency is ω = 21/6 ≃ 1.87083, and so the number of hertz is
ω/(2 π) ≃ .297752.
2
10.5.3. (a) Periodic of period π: -2 2 4 6 8 10

-1
-2
2
1.5
1
0.5
(c) Periodic of period 12: -5 5 15
10 20 25
-0.5
-1
-1.5
3
2
1
(f ) Quasi-periodic: 10 20 30 40 50 60
-1
-2
√ √
10.5.5. (a) 2 , 7 ; (b) 4 — each eigenvalue gives two linearly independent solutions;
! !
√ 2 √ −1
(c) u(t) = r1 cos( 2 t − δ1 ) + r2 cos( 7 t − δ2 ) ; (d) The solution
1 2
is periodic if only one frequency is excited, i.e., r1 = 0 or r2 = 0; all other solutions are
quasiperiodic.
√ √
10.5.7. (a) u(t) = r1 cos(t − δ1 ) + r2 cos( 5 t − δ2 ), v(t) = r1 cos(t − δ1 ) − r2 cos( 5 t − δ2 );
(c) u(t) = ( r1 cos(t − δ1 ), r2 cos(2 t − δ2 ), r3 cos(3 t − δ1 ) )T .
♠ 10.5.11. (a) The vibrational frequencies and eigenvectors are

q √ √ q √
ω1 = 2 − 2 = .7654, ω2 = 2 = 1.4142, , ω3 = 2+ 2 = 1.8478,
     
1 −1 1
√     √ 
v1 =  2 , v2 =  0  − 2 .
   , v3 =  
1 1 1
Thus, in the slowest mode, all three masses are moving in the same direction, with the mid-
√
dle mass moving 2 times farther; in the middle mode, the two outer masses are moving
in opposing directions by equal amounts, while the middle mass remains still; in the fastest
mode, the two outer masses are moving in tandem, while the middle mass is moving farther
in an opposing direction.

√ √
10.5.16. (a) u(t) = a t + b + 2 r cos( 5 t − δ), v(t) = − 2 a t − 2 b + r cos( 5 t − δ).
The unstable mode consists of the terms with a in them; it will not be excited if the initial
conditions satisfy u(t0 ) − 2 v(t0 ) = 0.

√ r √ ! √ r √ !
1− 13 7+ 13 1+ 13 7− 13
(c) u(t) = − 2 a t − 2 b − 4 r1 cos 2 t − δ1 − 4 r2 cos 2 t − δ2 ,
√ r √ ! √ r √ !
3− 13 7+ 13 3+ 13 7− 13
v(t) = − 2 a t − 2 b + 4 r1 cos 2 t − δ1 + 4 r2 cos 2 t − δ2 ,
r √ ! r √ !
7+ 13 7− 13
w(t) = a t + b + r1 cos 2 t − δ1 + r2 cos 2 t − δ2 .
The unstable mode is the term containing a; it will not be excited if the initial conditions
satisfy − 2 u(t0 ) − 2 v(t0 ) + w(t0 ) = 0.

 
− √1 √1 0 
4 0 0

 2 2 
  
10.5.17. (a) Q = 

0 0 1
,

Λ= 0
 2 0.
√1 √1 0 0 0 2
2 2
(b) Yes, because K is symmetric and has all positive eigenvalues.
T !
1√ √ √
(c) u(t) = cos2 t, √ sin 2 t, cos 2 t .
2
√
(d) The solution u(t) is periodic with period 2 π.
√
(e) No — since the frequencies 2, 2 are not rational multiples of each other,
the general solution is quasi-periodic.
r r
3 1
√ 3 1
√
♠ 10.5.22. (a) Frequencies: ω1 = 2 − 2 5 = .61803, ω2 = 1, ω3 = 2 5 = 1.618034;
+ 2
 √     √ 
2− 5 −1 2+ 5
     
 −1   −1   1 
stable eigenvectors: v1 = 
 √ 
, v2 =


,
 v3 = 
 √ 
; unstable
 −2 + 5   −1   −2 − 5 
1 1 1

1
 
 −1 
eigenvector: v4 =  
. In the lowest frequency mode, the nodes vibrate up and towards

 1
1
each other and then down and away, the horizontal motion being less pronounced than the
vertical; in the next mode, the nodes vibrate in the directions of the diagonal bars, with one
moving towards the support while the other moves away; in the highest frequency mode,
they vibrate up and away from each other and then down and towards, with the horizontal
motion significantly more than the vertical; in the unstable mode the left node moves down
and to the right, while the right-hand node moves at the same rate up and to the right.
! q !
1 −3
10.5.27. (a) u(t) = r1 cos √1 t − δ1 + r2 cos 5
t − δ2 ,
2 2 3 1
r √ !  1+√3  r √ !  1−√3 
3− 3  2 + 3+ 3  2 .
(c) u(t) = r1 cos 2 t − δ1 r2 cos 2 t − δ2
1 1
√ r √ √ r √
3−1 3− 3 3+1 3+ 3
10.5.28. u1 (t) = √
2 3
cos 2 t+ √
2 3
cos 2 t,
r √ r √
1 3− 3 1 3+ 3
u2 (t) = √ cos 2 t− √ cos 2 t.
2 3 2 3

♣ 10.5.30. (a) We place the oxygen molecule at the origin, one hydrogen at ( 1, 0 )T and the other
at ( cos θ, sin θ )T = ( −0.2588, 0.9659 )T with θ = 180
105
π = 1.8326 radians. There are two in-
dependent vibrational modes, whose fundamental frequencies are ω1 = 1.0386, ω2 = 1.0229,
with corresponding eigenvectors v1 = ( .0555, −.0426, −.7054, 0., −.1826, .6813 )T ,
v2 = ( −.0327, −.0426, .7061, 0., −.1827, .6820 )T . Thus, the (very slightly) higher frequency
mode has one hydrogen atoms moving towards and the other away from the oxygen, which
also slightly vibrates, and then all reversing their motion, while in the lower frequency mode,
they simultaneously move towards and then away from the oxygen atom.
10.5.37. The solution is u(t) = 41 (v+5) e− t − 41 (v+1) e− 5 t , where v = u(0) is the initial velocity.

v+1 v+1
This vanishes when e4 t = , which happens when t = t⋆ > 0 provided > 1, and
v+5 v+5
so the initial velocity must satisfy v < −5.
10.5.38. (a) u(t) = t e− 3 t ; critically damped. (c) u(t) = 1

4 sin 4 (t − 1); undamped.
− t/2 −t
(e) u(t) = 4 e − 2e ; overdamped.
10.5.39. (a) By Hooke’s Law, the spring stiffness is k = 16/6.4 = 2.5. The mass is 16/32 = .5.
√
The equations of motion are .5 u + 2.5 u = 0. The natural frequency is ω = 5 = 2.23607.

(b) The solution to the initial value problem .5 u + u + 2.5 u = 0, u(0) = 2, u(0) = 0, is

u(t) = e− t (2 cos 2 t + sin 2 t). (c) The system is underdamped, and the vibrations are less
rapid than the undamped system.
10.6.1. (a) cos 8 t − cos 9 t = 2 sin 12 t sin 17

2 t; fast frequency: 17
2 , beat frequency: 1
2.
2
-5 5 10 15 20
-1
-2
(c) cos 10 t + cos 9.5 t = 2 sin .25 t sin 9.75 t; fast frequency: 9.75, beat frequency: .25.
2
5 10 15 20 25 30
-1
-2
√ √ √
10.6.2. (a) u(t) = 1
27
1
cos 3 t− 27 cos 6 t, (c) u(t) = 1
2 sin 2 t+e− t/2 cos 15
2 t− 15
5 sin 15
2 t .

1 2 1
10.6.3. (a) u(t) = 3 cos 4 t + 3 cos 5 t + 5 sin 5 t;
undamped periodic motion with fast frequency 4.5 and beat frequency .5:
1
0.5
5 10 15 20 25 30
-0.5
-1
56 − 5 t
(c) u(t) = − 60
29 cos 2 t +
5
29 sin 2 t − 29 e + 8 e− t ;
the transient is an overdamped motion; the persistent motion is periodic:
4
3
2
1
-1 5 10 15 20 25
-2
10.6.4. In general, by (10.107), the maximal allowable amplitude is

q q
α= m2 (ω 2 − γ 2 )2 + β 2 γ 2 = 625 γ 4 − 49.9999 γ 2 + 1,
which, in the particular cases is (a) .0975, (b) .002.
10.6.6. The solution to .5 u + u + 2.5 u = 2 cos 2 t, u(0) = 2, u(0) = 0, is

u(t) = 4
17 cos 2 t + 16
17 sin 2 t + e− t 30
17 cos 2 t − 1
17 sin 2 t
= .9701 cos(2 t − 1.3258) + 1.7657 e− t cos(2 t + .0333).

The solution consists of a persistent periodic vibration at the forcing frequency of 2, with a
√
phase lag of tan−1 4 = 1.32582 and amplitude 4/ 17 = .97014, combined with a transient
vibration at the same frequency with exponentially decreasing amplitude.
10.6.10. (b) O verdamped, (c) critically damped, (d) underdamped.
3 − t/3 1 −t
10.6.11. (b) u(t) = 2e − 2e , (d) u(t) = e− t/5 cos 1
10 t + 2 e− t/5 sin 1
10 t.
165 − t/4
10.6.12. u(t) = 41 e cos 14 t− 91
41 e− t/4 sin 14 t − 124
41 cos 2 t + 32
41 sin 2 t
− .25 t − .25 t
= 4.0244 e cos .25 t − 2.2195 e sin .25 t − 3.0244 cos 2 t + .7805 sin 2 t.
10.6.14. (a) .02, (b) 2.8126.
10.6.17. u(t) =
1! q
√ ! √ !
2
√ −1 − 5 √ −1 + 5
(b) sin 3 t + r1 cos( 4 + 5 t − δ1 ) + r2 cos(4 − 5 t − δ2 ) ,
−1 2 2
     
2 q
1  17  −3 1
(d) cos 2t + r1 cos( 53 t − δ1 )  + r cos( √1
2 t − δ2 ) .
2
− 12
17 1 2

r √ r√
3− 3 3+ 3
10.6.18. (a) The resonant frequencies are 2 = .796225, 2 = 1.53819.
r √ ! !
3+ 3 w1
(b) For example, a forcing function of the form cos 2 t w, where w = is not
w2
√ !
−1 − 3 √
orthogonal to the eigenvector , so w2 6= (1 + 3 )w1 , will excite resonance.
1
♣ 10.6.20. In each case, you need to force the system by cos(ω t)a where ω 2 = λ is an eigenvalue
and a is orthogonal to the corresponding eigenvector. In order not to excite an instability,
a needs to also be orthogonal to the kernel of the stiffness matrix spanned by the unstable
mode vectors.
(a) Resonant frequencies: ω1 = .5412, ω2 = 1.1371, ω3 = 1.3066, ω4 = 1.6453;
       
.6533 .2706 .2706 − .6533
       
 .2706   .6533   − .6533   .2706 
eigenvectors: v1 = 
, v = 
 2 
, v = 
 3 
, v = 
 4 
;

 .6533   − .2706   .2706   .6533 
− .2706 .6533 .6533 .2706
no unstable modes.
(c) Resonant frequencies: ω1 = .3542, ω2 = .9727, ω3 = 1.0279, ω4 = 1.6894, ω5 = 1.7372;
eigenvectors:
         
− .0989 − .1160 .1251 .3914 .6889
         
 − .0706   .6780   − .6940   .2009   .1158 
         
         


0  
 .2319 
 


0  
 − .7829 
 


0  ;
v1 =  , v2 =  , v3 =  , v4 =  , v5 =  
 − .9851   0   .0744   0   − .1549 
         
 .0989   − .1160   − .1251   .3914   − .6889 
         
− .0706 − .6780 − .6940 − .2009 .1158
unstable mode: z = ( 1, 0, 1, 0, 1, 0 )T . To avoid exciting the unstable mode, the initial veloc-
ity must be orthogonal to the null eigenvector: z · u(t0 ) = 0, i.e., there is no net horizontal

velocity of the masses.

http://www.springer.com/978-3-319-91040-6

Ala2 SSM PDF

Uploaded by

Copyright:

Available Formats

Ala2 SSM PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ala2 SSM PDF

Uploaded by

Copyright:

Available Formats

Students’ Solutions

Applied Linear Algebra

Using the Manual

Peter J. Olver Cheri Shakiban

Minnesota, September 2018

Chapter 1. Linear Algebraic Systems . . . . . . . . . . . . . . . . . 1

Chapter 2. Vector Spaces and Bases . . . . . . . . . . . . . . . . 11

Chapter 3. Inner Products and Norms . . . . . . . . . . . . . . . 19

Chapter 5. Minimization and Least Squares . . . . . . . . . . . . . 36

Chapter 8. Eigenvalues and Singular Values . . . . . . . . . . . . . 58

Chapter 10. Dynamics . . . . . . . . . . . . . . . . . . . . . . 81

1.1.1. (b) Reduce the system to 6 u + v = 5, − 25 v = 5

1.2.5. (b) u + w = −1, u + v = −1, v + w = 2. The solution is u = −2, v = 1, w = 1.

1.2.9. 1, 6, 11, 16.

1.2.11. (a) True.

1.2.17. A On×p = Om×p , Ol×m A = Ol×n .

1.2.22. (a) A must be a square matrix. (b) By associativity, A A2 = A A A = A2 A = A3 .

1.3.3. (a) 3 x + 2 y = 2, − 4 x − 3 y = −1; solution: x = 4, y = −5.

1.4.1. (a) Nonsingular, (c) nonsingular, (e) singular.

1.4.2. (a) Regular and nonsingular, (c) nonsingular.

1.4.5. (b) x1 = 0, x2 = −1, x3 = 2; (d) x = − 13 9

1.4.6. True. All regular matrices are nonsingular.

♦ 1.5.19. (a) A = I −1 A I . (b) If B = S −1 AS, then A = S B S −1 = T −1 B T , where T = S −1 .

♦ 1.6.13. (a) Using Exercise 1.6.12, aij = eT T

1.6.17. (b) a = −1, b = 2, c = 3.

♦ 1.6.30. Write A = L D V , then AT = V T D U T = L eU e = V T and U

♦ 1.7.4. We begin by proving (1.63). We must show that 1 + 2 + 3 + . . . + (n − 1) = 21 n(n − 1)

1.7.11. For n = 4: (a)

1.8.1. (a) Unique solution: (− 21 , − 34 )T ; (c) no solutions;

1.8.2. (b) Incompatible; (c) (1, 0)T ;

1.8.7. (b) 1, (d) 3, (f ) 1.

1.8.11. Example: (a) x2 + y 2 = 1, x2 − y 2 = 2.

1.8.22. (a) x = z, y = z, where z is arbitrary; (c) x = y = z = 0.

1.9.5. By (1.85, 86) and commutativity of numeric multiplication,

♥ 1.9.20. (a) By direct substitution:

2.1.1. Commutativity of Addition:

Additive Identity: 0 = 0 = 0 + i 0 and

2.1.4. (a) ( 1, 1, 1, 1 )T , ( 1, −1, 1, −1 )T , ( 1, 1, 1, 1 )T , ( 1, −1, 1, −1 )T . (b) Obviously not.

2.1.6. (a) f (x) = − 4 x + 3.

2.1.11. (j) Let z = c 0. Then z + z = c (0 + 0) = c 0 = z, and so, as in the proof of (h), z = 0.

2.2.3. (a) Subspace: 0

2.2.7. (b) Not a subspace; (d) subspace.

2.2.10. (a) Not a subspace; (c) not a subspace; (f ) subspace;.

2.2.14. (a) Vector space; (c) vector space;

2.2.15. (b) Subspace; (d) subspace, (e) not a subspace.

2.2.20. ∇ · (c v + d w) = c ∇ · v + d ∇ · w = 0 whenever ∇ · v = ∇ · w = 0 and c, d, ∈ R.

♦ 2.2.22. (a) If v, w ∈ W ∩ Z, then v, w ∈ W , so c v + d w ∈ W because W is a subspace, and

♦ 2.2.30. (a) By induction, we can show that, for n ≥ 1 and x > 0,

2.3.4. (a) No, (c) yes, (e) yes.

2.3.9. (a) Yes, (c) no, (e) no.

2.3.21. (b) Linearly dependent; (d) linearly independent; (f ) linearly dependent.

2.3.27. Yes, when it is the zero vector.

♦ 2.3.31. (a) Since v1 , . . . , vn are linearly independent,

2.3.33. (b) Linearly independent; (d) linearly independent; (f ) linearly dependent.

♥ 2.3.37. (a) If c1 f1 (x) + · · · + cn fn (x) ≡ 0, then c1 f1 (xi ) + · · · + cn fn (xi ) = 0 at all sample

2.4.1. (a) Basis; (b) not a basis (d) not a basis.

2.4.2. (a) Not a basis; (c) not a basis.

2.4.8. (b) The condition p(1) = 0 says a + b + c = 0, so

p(x) = (− b − c) x2 + b x + c = b (− x2 + x) + c(− x2 + 1).