MAST10018 Exercises

MAST10018
Linear Algebra
Extension Studies
Exercises
Department of Mathematics and Statistics
© University of Melbourne 2016

This compilation has been made in accordance with the
provisions of Part VB of the Copyright Act (1968) for the teaching purposes of
the University of Melbourne. No part of this publication may be reproduced or
transmitted by any form, except as permitted under this act.
Content last updated: 4 January 2018
MAST10018: Linear Algebra Extension Studies Exercises
Contents
Exercises 1: Linear equations and matrices 2
Exercises 2: Vectors and geometry 7
Exercises 3: Vector spaces 11
Exercises 4: Inner product spaces 18
Exercises 5: Linear transformations 22
Exercises 6: Eigenvalues and eigenvectors 28
Answers 1: Linear equations and matrices 31
Answers 2: Vectors and geometry 39
Answers 3: Vector spaces 44
Answers 4: Inner product spaces 50
Answers 5: Linear transformations 55
Answers 6: Eigenvalues and eigenvectors 61
Content last updated: January 4, 2018
1
MAST10018: Linear Algebra Extension Studies Exercises 1: Linear equations and matrices
Exercises 1: Linear equations and matrices
Arithmetic of matrices
1. Evaluate the following matrix products:

  
3 4 2 1 2  
   2 2 h i
(a)  1 3 6   0 −1  (b)  −1 −2
   
   1 −1
7 −1 0 −1 0
   
0 0
h i   h i
(c) −4 + 3i  1−i (d)  1  7 −4
   
7+i 6  6
   
1 1
    
3 −6 0 2 4 3  
     1
2
(e)  0 2 −2   1  (f)  6
    
6  
     1
3
1 −1 −1 1 8 9
2. Let  
1 0 1
A= 2 3 4 .
−1 0 −2
Show that A−1 = − 31 (A2 − 2A − 4I) where A−1 is a matrix such that A−1 A = AA−1 = I.
3. Verify that

cos (θ1 ) − sin (θ1 ) cos (θ2 ) − sin (θ2 ) cos (θ1 + θ2 ) − sin (θ1 + θ2 )
= .
sin (θ1 ) cos (θ1 ) sin (θ2 ) cos (θ2 ) sin (θ1 + θ2 ) cos (θ1 + θ2 )
4. Write down the 3 × 2 matrices A and B that have entries given by Aij = i + j and Bij = (−1)i+j .
Calculate AT B.
5. Give 3 × 3 examples of the following:
(a) a diagonal matrix; that is, a matrix A with Aij = 0 if i 6= j.
(b) a scalar matrix; that is, a diagonal matrix A with Aii = Ajj .
(c) a symmetric matrix; that is, a matrix A with Aij = Aji .
(d) an upper triangular matrix; that is, a matrix A with Aij = 0 if i > j.
6. Use trial and error to find 2 × 2 examples of the following:
(a) a non-zero matrix A with A2 = 0;
(b) a matrix B with real entries and with B 2 = −I;
(c) matrices C and D with no zero entries but with CD = 0;
(d) matrices E and F with EF = −F E but EF 6= 0.
2
7. Suppose that a 2 × 2 matrix A satisfies AB = BA for every 2 × 2 matrix B. Set

a b
A= .
c d
Then AB = BA when B is either of

1 0 0 1
or .
0 0 0 0
What does this say about a, b, c, d? Show that a matrix A that satisfies the above must be a scalar
matrix.
8. Let A be a square matrix satisfying A2 = A and let B be any matrix of the same size. Show that
(AB − ABA)2 = 0
Linear equations
9. Which of the following systems of equations are linear?
(a) x1 − 3x2 = x3 − 4 (b) 3x − xy =1

x4 = 1 − x1 x + 2xy − y = 0
x1 + x4 + x3 − 2 = 0
(c) y = x − 1 (d) x2 = y
x=1−y x+y =1
The row reduction algorithm
10. Following the algorithm, reduce the following matrices to row echelon (RE) form, and then to reduced
row echelon (RRE) form. Keep a precise record of the elementary row operations you use.
 
4 −8 16  
  2+i 2+i 5 6+i
(a)  1 −3 (b) 
 
6  
  1 − 2i 1 − 2i −2 + i 2−i
2 1 1
 
1 2  
  0 2 1 4
 −1
 
1   
(c)  (d)  0
 


 0 2 6 
 2 2   
  1 0 −3 2
0 2
 
  0 0 2 7
1 2 0 1  
−1
 
   1 1 1 
(e)  2 (f) 
 
4 1 1  
 −1 1 −4
 
  5 
3 6 1 1  
−2 2 −5 4
Homogeneous linear equations
11. For each of the matrices of Q10 write down the corresponding homogeneous system, the rank of the
matrix (where rank is the number of leading entries or non-zero rows in RE form) and solve this system of
equations from the RRE form. N.B. Elementary row operations leave columns with all zeros unchanged.
12. What is the rank of the n × n matrix with 1 in every position? What is the rank of the ‘chessboard matrix’
with (i, j) entry ai,j satisfying ai,j = 0 when i + j is even and ai,j = 1 when i + j is odd?
3
13. Using row reduction, solve the following homogeneous systems:
(a) (2 + i)x − iy + (3 − i)z + u = 0 (b) 7x + 7y − 16z + 6w = 0

x+ y − (1 − i)u = 0 4x − y + 3z + 7w = 0
3x + y − 2z + 4w = 0
Inhomogeneous linear equations
14. Use row reduction to find the ranks of the coefficient and augmented matrices of the following systems of
equations. Decide whether the system has (a) no solution vector, (b) a unique solution vector, (c) more
than one solution vector (x, y, z). Solve the system where possible and describe it geometrically.
(a) 3x − 2y + 4z = 3 (b) x + 2y − z = −1
x− y+ z = 7 2x + 7y − z = 3
4x − 3y + 5z = 1 −3x − 12y + z = 0
(c) 3x − 4y + z = 2 (d) 2x − 3y + 5z = 10
−5x + 6y + 10z = 7 4x + 7y − 2z = −5
8x − 10y − 9z = −5 2x − 4y + 25z = 31
15. Using row reduction, find the general solution to the following system of equations:
2x1 + x2 + 3x3 + x4 = 3
x1 + x2 + x3 − x4 = 6
x1 − x2 + 3x3 + 5x4 = −12
4x1 + x2 + 7x3 + 5x4 = −3
16. Determine the values of k for which the system of linear equations has (i) no solution vector, (ii) a unique
solution vector, (iii) more than one solution vector (x, y, z):
(a) kx + y + z = 1 (b) 2x + (k − 4)y + (3 − k)z = 1

x + ky + z = 1 2y + (k − 3)z = 2
x + y + kz = 1 x− 2y + z = 1
(c) x + 2y + kz = 1 (d) x − 3z = −3
2x + ky + 8z = 3 2x + ky − z = −2
x + 2y + kz = 1
For the cases (ii) and (iii) find the solutions.
17. Determine the conditions on a, b, c so that the system has a solution:
(a) x + 2y − 3z = a (b) x − 2y + 4z = a
3x − y + 2z = b 2x + 3y − z = b
x − 5y + 8z = c 3x + y + 2z = c
Find the solutions when they exist.
4
Inverses
18. Use elementary row operations to find, where possible, the inverses of the following matrices:
   
0 0 1 1 0 0
   
(a)  0 1 0  (b)  0
   
4 3 
   
1 0 0 0 9 7
   
4 7 2 1 2 3
   
(c)  −3 −7  (d)  −1 7 −4 
   
1
   
2 4 1 0 9 −1
19. Repeat the preceding problem for the complex matrices

   
1+i i 2 −i
(a)   (b)  
1 1−i 3+i 1
20. Find the inverse of the matrix  

4 3 −2 0
 1 2 −3 1 
A= .
 2 −1 1 −3 
1 −3 1 −2
Hence solve the system
4x1 + 3x2 − 2x3 = 1
x1 + 2x2 − 3x3 + x4 = 0
2x1 − x2 + x3 − 3x4 = −1
x1 − 3x2 + x3 − 2x4 = 3
Check your answer by matrix multiplication.
Determinants
21. Evaluate the following determinants by reduction to triangular form:

−1 −2

2 3 1 4 5 6 2

(a) 4 (b) 1 2 (c) −1

5 2 3 2 1

0 1 −1 0 −3

1 2 3 3

−1

1 2 3 4 1 2 3 3 9 27 81

0 1 2 3 2 1 2 6 1 1 1 1
(d) (e) (f)
−2 4 −8

0 0 2 1 1 0 2 3 16

−2 0 −5

0 0 3 2 2 2 4 8 16
5
22. Let
a b c

|A| = d e f = 1.

g h i
Find the following determinants:

−b

a b c a c d e f

(a) g (b) d −e f (c) 3g

h i 3h 3i

−h i

d e f g a b c

2a 2b 2c a b c

(d) 2d (e) d + a

2e 2f e+b f +c

g − 2a h − 2b i − 2c

2g 2h 2i
23. Using a suitable determinant find
(a) the area of the parallelogram spanned by the vectors (1, 2) and (−3, 5),
(b) the volume of the parallelepiped spanned by the vectors (7, 1, 1), (4, 1, 2) and (1, 0, −1)
(c) the volume of the parallelepiped spanned by the vectors (1, 1, 1), (4, 1, 2) and (2, −1, 0)
What does your answer to (c) tell you about the three vectors?
24. Establish the following factorizations:

1 1 1

a b c
3 3 3 = (b − c)(c − a)(a − b)(a + b + c)
a b c

x
2 y2 z2

x
y z = (y − z)(z − x)(x − y)(yz + zx + xy)
yz zx xy
25. Evaluate the following determinants:

1
a bc 1
a a3

(a) 1 b ca (b) 1
b b3

1 c ab 1 c c3
26. The famous Vandermonde determinant is defined by

1 λ1 λ21 λn−1

··· 1

λn−1

1 λ2 λ22 ···
2
Dn = .

.. .. .. ..
. . .

1 λn λ2 ··· λn−1
n n
Use row operations to evaluate D2 , D3 , and D4 , expressing your answer as a fully factorized expression
to exhibit the fact that Dn = 0 if any two of the parameters λ1 , λ2 , . . ., λn are equal.
27. Idempotent Matrices A matrix P is called idempotent if P 2 = P . If P is idempotent and P 6= I show

that det P = 0.
6
MAST10018: Linear Algebra Extension Studies Exercises 2: Vectors and geometry
Exercises 2: Vectors and geometry

In the following, vectors are denoted by bold type and e1 , e2 , e3 denote (1, 0, 0), (0, 1, 0), (0, 0, 1) respectively.
Arithmetic of vectors
1. For the given vectors a, b find a + b, 5a − 4b, a · b, kak, and ka − bk.
(a) a = (−2, 6, 1), b = (3, −3, −1)
(b) a = 3e1 − 4e2 + 21e3 , b = e1 + 2e2 − 5e3
(c) a = e1 + e2 , b = −e2 + e3
2. Find the following scalar products:

√ √
(a) (1, 1, 1) · (2, 1, −3) (b) (2, 1, 1) · (1, −3, 7) (c) ( 2, π, 1) · ( 2, −2, 3)
3. Find the angle between the following pairs of vectors:
(a) (1, 0, 0), (0, 0, 4) (b) (1, −1, 0), (0, 1, 1) (c) (2, −2, 2), (−1, 0, 2)
4. If a = 2e1 + xe2 + e3 and b = 4e1 − 2e2 − 2e3 , find x such that a is orthogonal to b. Can you find a value
of x so that a and b are in the same direction?
5. Find the following vector products:
(a) (e1 − e2 − 2e3 ) × (2e1 + e2 − 3e3 ) (b) (2, 1, 1) × (1, −3, 7)
6. Find unit vectors orthogonal to the following pairs of vectors:
(a) (2, 0, 1), (3, −1, 4) (b) (2, 5, 3), (−1, −5, 3) (c) (5, 4, 2), (3, 2, 1)
7. Find the area of a triangle with the vertices (1, −1, 2), (−2, 1, 1), (1, 2, 3). Find a unit vector orthogonal
to the plane of this triangle.
8. Find the volume of a parallelepiped with the sides a = (1, −2, 1), b = (2, 3, −1), c = (−4, 2, 3).
Some vector identities
9. If a · b = 0 and x + (x · b)a = b, find the vector x in terms of a and b.
10. Show that a · (b × c) = c · (a × b) = −b · (a × c).
11. Prove that a × (b × c) = (a · c)b − (a · b)c.

P3
Hint. Find ek × (b × c) for k = 1, 2, 3. Express a × (b × c) as a sum k=1 ak ek × (b × c).
12. Use Qs 10 and 11 to show that (a × b) · (c × d) = (a · c) (b · d) − (a · d) (b · c).
Deduce Lagrange’s identity: ka × bk2 = kak2 kbk2 − (a · b)2 .
7
13. Prove the following vector identities:
(a) a × (b × c) + b × (c × a) + c × (a × b) = 0
(b) (b × c) · (a × d) + (c × a) · (b × d) + (a × b) · (c × d) = 0
Hint. Use Qs 11 and 12.
14. Prove the following properties of the vector product:
(a) a × b = −b × a (b) (αa + βb) × c = αa × c + βb × c

(c) a × (αb + βc) = αa × b + βa × c (d) a × a = 0
Planes in R3
15. Find the orthogonal projection of the vector x on the vector a:
(a) x = (3, −4, 5), a = (−3, 1, 2)
(b) x = (12, −17, 8), a = (1, −1, 2).
16. Find a Cartesian equation of the plane
(a) through the points (2, 0, 1), (1, 1, 3), (−5, 2, 3)
(b) through the point (2, 0, 1) orthogonal to the vector (1, 2, 3).
17. Find the distance of the point (1, 1, 1) from the plane 2x + y + 2z − 2 = 0.
18. Find the angle between the planes 2x + y + 3z = 0 and 3x − 2y + 4z − 4 = 0.
Lines in R3
19. Find a parametric equation of the line `:
(a) ` is the line through P (3, 4, −7) and parallel to a = (1, 2, −3)
(b) ` is the line through the points A(2, 0, 1) and B(3, −4, 1).
20. Find the angle between the lines
(a) x − 3 = 2 − y, z = 1, and x = −3, y + 2 = z − 5
x−2 y+5 z−1

(b) = = and x = −y = z .
3 4 3
21. Find the distance of the point A from the line `:
(a) A(6, 1, 2), and ` : 2 − x = z − 2, y = 2
(b) A(3, −1, −1), and ` is the line through the points P (1, 0, −2) and Q(2, 1, −1).
22. Find the distance between the lines
(a) x − 3 = 4y − 2 = z + 1 and x = 3 − 2t, y = t, z = 1 + t
x−3 z x y+1 z−2

(b) =y+2= and = =
2 8 3 2 13
8
23. Find parametric equations of the line of intersection of the planes 2x − y − z = 0 and x + y + 5 = 0.
24. Find the distance of the point (1, 1, 1) from the line of intersection of the planes x + y + z = 0 and
x − 2y − 3 = 0.
25. Find the angle between the line x = 2t − 7, y = 4t − 6, 3z = t − 5 and the plane
x = −2s, y = 2t + s, z = t − s
Miscellaneous problems on lines and planes
26. Find the intersection of the line 2x = 3y = −z with the plane x + 4y + 3z + 14 = 0.
27. Find a Cartesian equation of the plane through the points (2, −1, 0) and (−5, −3, 1) which is parallel to
the line joining the points (3, 5, −1) and (0, 3, −2). Write down a parametric equation of the line through
the origin which is orthogonal to this plane, and find where it meets the plane.
28. Find the foot P of the perpendicular from the point A(−2, 3, 1) to the plane π with equation 2x−y+z−1 =
0. Check by another method that the distance from A to π is equal to the distance from A to P .
29. (Exam, 1988) If the line ` is given by the equations 2x − y + z = 0, x + z − 1 = 0, and if M is the point
(1, 3, −2), find a Cartesian equation of the plane
(a) passing through M and `,
(b) passing through M and orthogonal to `.
30. (Hard) The gutters on two roofs meet at right angles and the roofs themselves make angles of α and β
with the horizontal. Show that the line of intersection of the roofs makes an angle with the horizontal of
!
1
arcsin p .
1 + cot2 (α) + cot2 (β)
31. (Hard) The legs of a tripod are at right angles to each other. Two legs have the same length a and the
third is longer with length b. The lower ends of the legs are placed on a level floor. What is the height of
the top of the tripod above the floor?
32. A median of a triangle is the line connecting a vertex to the midpoint of the opposite side. Use vector
methods to show that the three medians of a triangle intersect in a point.
33. An altitude of a triangle is the line through a vertex which is perpendicular to the opposite side. Use
vector methods to show that the three altitudes of a triangle intersect in a point.
34. (Hard) Use vector methods to prove Desargues Theorem:
Two triangles ABC and A0 B 0 C 0 in space are situated so that the lines AA0 , BB 0 and CC 0 meet in a point
O. Let
P denote the intersection of BC with B 0 C 0
Q the intersection of CA with C 0 A0
R the intersection of AB with A0 B 0 .
Assuming that all these intersections exist, show that P , Q and R lie on a straight line.
What can be said if, for example, BC is parallel to B 0 C 0 ?
9
35. (Hard) Use vector methods to prove Pappus Theorem:
Points A, B, C lie on a line ` and points A0 , B 0 , C 0 lie on a line `0 which meets `. Let
P denote the intersection of BC 0 and B 0 C
Q denote the intersection of CA0 and C 0 A
R denote the intersection of AB 0 and A0 B.
Assuming that all these intersections exist, show that P , Q and R lie on a straight line.
What can be said if, for example, BC 0 is parallel to B 0 C?
36. (Hard) Use vector methods to prove the theorem of Menelaus:

A line cuts the sides AB, BC and CA of a triangle ABC (with the sides extended beyond the triangle)
in the points L, M and N , respectively (see the diagram for an example). Prove that
AL BM CN
· · = −1
LB M C N A
where the lengths of the segments are given appropriate signs.
10
MAST10018: Linear Algebra Extension Studies Exercises 3: Vector spaces
Exercises 3: Vector spaces

Linear combinations and linear dependence, AR5.3
1. Let u = (1, 0, −1) and v = (−2, 1, 1).
(a) Write w1 = (−1, 2, −1) as a linear combination of u and v.
(b) Show that w2 = (−1, 1, 1) cannot be written as a linear combination of u and v.
(c) For what value of c is the vector (1, 1, c) a linear combination of u and v?
2. In this question let S = {v1 , v2 , v3 , v4 , v5 } be vectors in R3 and let A be the 3 × 5 matrix with the ith
column given by the vector vi . Suppose that the RRE form of A is
 
1 2 0 −1 0
0 0 1 3 0
0 0 0 0 1
Are the following sets linearly dependent or independent? If linearly dependent, express one vector as a
linear combination of the others.
(a) {v1 , v2 , v3 } (b) {v1 , v3 , v4 } (c) {v1 , v4 , v5 } (d) {v3 , v4 , v5 }
3. Determine whether or not the following sets of vectors are linearly independent:
(a) {(2, −3, 1, −5), (0, 1, 2, 2), (1, −2, 3, 0)}
(b) {(1, 0, 2, −3), (0, −4, 1, 1), (2, 2, 0, −1), (1, −2, −1, 3)}
4. Determine whether the following sets are linearly dependent or linearly independent.
(a) {(1, 2), (0, 2), (1, 0), (−1, 1)}
(b) {(1, 2), (3, −1)}
(c) {(1, 0, 1), (−1, 1, 0), (0, 1, 1))}
(d) {(2, 0, 0, 0), (2, 1, 0, 0), (−1, 3, −2, 0), (1, −2, 4, −3)}
5. Which of the following sets of vectors in C3 are linearly independent?
(a) {(1 − i, 1, 0), (2, 1 + i, 0), (1 + i, i, 0)}
(b) {(1, 0, −i), (1 + i, 1, 1 − 2i), (0, i, 2)}
(c) {(i, 0, 2 − i), (0, 1, i), (−i, −1 − 4i, 3)}
6. (Hard) Show that the vectors (1, a, a2 ), (1, b, b2 ), (1, c, c2 ) are linearly independent if a, b, c are all distinct
(i.e., a 6= b, a 6= c and b 6= c). Can you generalise this?
Subspaces of Rn , AR5.2
7. For each of the following subsets of R2 sketch the set, then determine whether it is (i) closed under
addition, (ii) closed under scalar multiplication, (iii) a subspace of R2 .
(a) {(x, y) : y > 0} (b) {(x, y) : x = y}
(c) {(x, y) : x2 + y 2 6 1} (d) {(x, y) : xy = 0}.
11
8. Decide which of the following are subspaces of R3 . Explain your answers.
(a) {(a, b, 0) ∈ R3 : a, b ∈ R} (b) {(a, b, c) ∈ R3 : 2a − 3b + 5c = 4}

(c) {(a, b, c) ∈ R3 : 2a − 3b + 5c = 0} (d) {(a1 , a2 , a3 ) ∈ R3 : a1 > 0}
(e) {(a − b, a + b, 2a) ∈ R3 : a, b ∈ R}
9. Show that the following sets of vectors are subspaces of Rm .
(a) The set of all linear combinations of the vectors (1, 0, 1, 0) and (0, 1, 0, 1) (of R4 ).
(b) The set of all vectors of the form (a, b, a − b, a + b) (of R4 ).
(c) The set of all vectors (x, y, z) such that x + y + z = 0 (of R3 ).
10. Show that the following sets of vectors are not subspaces of Rm .
(a) The set of all vectors whose first component is 2.
(b) The set of all vectors except the vector 0 .
(c) The set of all vectors the sum of whose components is 1.
Spanning sets and bases, AR5.2, AR5.4
11. Determine whether the given set spans the given vector space.

1 3
(a) In R2 : , .
2 4
       
 2 3 1 7 
(b) In R3 :  0  ,  1  ,  1  ,  3  .
1 2 1 5
 
12. Determine which of the following sets span R3 .
(a) {(1, 2, 3), (−1, 0, 1), (0, 1, 2)}
(b) {(−1, 1, 2), (3, 3, 1), (1, 2, 2)}
13. Find spanning sets for the following subspaces of R3 :
(a) {(2a, b, 0) : a, b ∈ R}
(b) {(a + c, c − b, 3c) : a, b, c ∈ R}
(c) {(4a + d, a + 2b, c − b) : a, b, c, d ∈ R}
14. Find a set that spans the subspaces given in Q9.
15. In each part determine whether or not the given set forms a basis for the indicated (sub)space.
(a) {(1, 2, 3), (−1, 0, 1), (0, 1, 2)} for R3
(b) {(−1, 1, 2), (3, 3, 1), (1, 2, 2)} for R3
(c) {(1, −1, 0), (0, 1, −1)} for the subspace of R3 consisting of all (x, y, z) such that x + y + z = 0.
(d) {(1, 1, 0), (1, 1, 1)} for the subspace of R3 consisting of all (x, y, z) such that y = x + z.
12
16. Which of the following sets of vectors are bases for R3 ?
(a) {(1, 0, 0), (2, 2, 0), (3, 3, 3)}
(b) {(2, −3, 1), (4, 1, 1), (0, −7, 1)}
17. Find a basis for and the dimension of the subspace of Rn spanned by the following sets.
(a) {(0, 1, −2), (3, 0, 1), (3, 2, −3)} (n = 3)
(b) {(1, 3), (−1, 2), (7, 6)} (n = 2)
(c) {(−1, 2, 0, 4), (3, 1, −1, 2), (−5, 3, 1, 6), (7, 0, −2, 0)} (n = 4)
18. For each of the following sets choose a subset that is a basis for the subspace spanned by the set. Then
express each vector that is not in the basis as a linear combination of the basis vectors.
(a) (1, 2, 0, −1), (2, −1, 2, 3), (−1, −11, 6, 13), (4, 3, 2, 1)
(b) (0, −1, −3, 3), (−1, −1, −3, 2), (3, 1, 3, 0), (0, −1, −2, 1)
(c) (1, 2, −1), (0, 3, 4), (2, 1, −6), (0, 0, 2)
[Suggestion: Write down the given vectors as columns.]
19. In each part explain why the given statement is true “by inspection.”
(a) The set {(1, 0, 3), (−1, 1, 0), (1, 2, 4), (0, −1, −2)} is linearly dependent.
(b) The set {(1, −1, 2), (0, 1, 1)} does not span R3 .
(c) If the set {v1 , v2 , v3 , v4 } of vectors in R4 is linearly independent, then it spans R4 .
(d) The set {(0, 1, −1, 0), (0, −1, 2, 0)} is linearly independent, and so it spans the subspace of R4 of all
vectors of the form (0, a, b, 0).
20. In each part find a basis for and the dimension of the indicated subspace.
(a) The solution space of the homogeneous linear system:
x1 − 2x2 + x3 = 0
x2 − x3 + x4 = 0
x1 − x2 + x4 = 0
(b) The solution space of
x1 − 3x2 + x3 − x5 = 0
x1 − 2x2 + x3 − x4 = 0
x1 − x2 + x3 − 2x4 + x5 = 0
(c) The subspace of R4 of all vectors of the form (x, −y, x − 2y, 3y).
13
Row, column and solution spaces, AR5.5
21. Explain why the dimensions of the row space and the column space of a matrix are the same. What are
they for the following matrices?
 
1 −1 3
     
1 0 −1
 
1 2 1 0 1 1
(a)   (b)   (c)  
2 1 −1 −1 0 1
 
1 1 0
 
2 −1 1
22. Find a basis for the column space for each matrix of Q21.
23. Find a basis for the row space for each matrix of Q21.
24. Find a basis for the solution space for each matrix of Q21.
25. Find bases for the following subspaces of R3 .
(a) The set of vectors lying in the plane 2x − y − z = 0.
(b) The set of vectors on the line x/2 = y/3 = z/4.
General vector spaces, AR5.1
26. (Function spaces) Show that the set F(S, F) of all functions f : S → F is a vector space under the
pointwise operations
(f + g)(s) = f (s) + g(s), (λf )(s) = λf (s)
for all s ∈ S taking the scalar field F to be R or C.
Note. Most of the real and complex vector spaces encountered in this course are subspaces of Rn , Cn or
F(S, F) for some S and F = R or C.
27. Prove the following consequences of the vector space axioms:
(a) λ0 = 0; (b) 0x = 0; (c) (−1)x = −x.
28. Use the subspace theorem to decide which of the following are real vector spaces with the usual operations.
(a) The set of real polynomials of any degree.
(b) The set of real polynomials of degree 6 n.
(c) The set of real polynomials of degree exactly n.
(d) The set of real polynomials p with p(0) = 0.
(e) The set of real polynomials p with p(0) = 1.
(f) The set of all differentiable functions.
(g) The set of all solutions of the differential equation y 00 − 3y 0 + 2y = 0.
14
29. Determine whether or not the given set is a vector space under the usual operations. If it is not a vector
space, list all properties that fail to hold.
(a) The set of all 2 × 3 matrices whose second column consists of 0’s.
(b) The set of all (real) polynomials with strictly positive coefficients.
(c) The set of all (real valued) continuous functions with the property that the function is 0 at every
integer, for example f (x) = sin(πx).
30. Determine whether or not the given set is a subspace of M2,2 :
(a) The set of all 2 × 2 matrices, the sum of whose entries is zero.
(b) The set of all 2 × 2 matrices whose determinant is zero.
31. Determine whether or not the given set is a subspace of Mn,n , the space of all square matrices of size n.
(a) The diagonal matrices of order n.
(b) The matrices of order n with trace equal to 0. (The trace of a square matrix is the sum of the
diagonal elements.)
32. (Intersections of subspaces) Let H and K be subspaces of a vector space V . Prove that the intersection
K ∩ H is a subspace of V .
Linear dependence, spanning sets, bases, AR5.2, 5.3, 5.4

6 −8 4 0 1 −1 0 2
33. Is a linear combination of the matrices A = ,B = , and C = ?
−1 −8 −2 −2 2 3 1 4
34. Express −9−7x−15x2 as a linear combination of p1 = 2+x+4x2 , p2 = 1−x+3x2 , and p3 = 3+2x+5x2 .
35. Determine whether or not the given set is linearly independent. If the set is linearly dependent, write one
of its vectors as a linear combination of the others.
(a) {1, 1 + x, 1 + x + x2 } in P2
(b) {1 + x2 , 1 + x + 2x2 , x + x2 } in P2

1 0 0 1 0 0 0 0
(c) , , , in M2,2
0 −1 0 −1 1 −1 0 1
(d) {1, sin2 x, cos2 x} in C(−∞, ∞), the vector space of all continuous functions from R to R.
   
 2 0 0 −2 0 0 
(e) 0 −1 0 ,  0 −1 0  in M3,3
0 0 1 0 0 −1
 
36. Determine whether or not each set in Q35 generates (that is, spans) the indicated vector space.
37. Determine whether or not each set in Q35 is a basis for the indicated vector space.
15
38. Find the dimension of the given vector space:
(a) The subspace of M2,2 consisting of all diagonal 2 × 2 matrices.
(b) The subspace of M2,2 consisting of all 2 × 2 matrices whose diagonal entries are zero.
(c) The subspace of P3 consisting of all polynomials p(x) = a0 + a1 x + a2 x2 + a3 x3 with a2 = 0.
39. Determine whether the given set of vectors spans the given vector space.
(a) In P2 : {1 − x, 3 − x2 }.

2 1 0 0 3 −1 0 0
(b) In M2,2 : , , , .
0 0 2 1 0 0 3 1
40. Which of the following sets of vectors are bases for P2 ?
(a) {1 − 3x + 2x2 , 1 + x + 4x2 , 1 − 7x} (b) {1 + x + x2 , x + x2 , x2 }
41. Show that the following set of vectors is a basis for M2,2 :

3 6 0 −1 0 −8 1 0
, , , .
3 −6 −1 0 −12 −4 −1 2
42. (a) Show that any four polynomials in P2 are linearly dependent.
(b) Show that two polynomials cannot span P2 .
(c) (Hard) Prove that if V and W are three-dimensional subspaces of R5 , then V and W must have a
nonzero vector in common. (Hint: Start with bases for the two subspaces, making six vectors in all.)
43. (Symmetric matrices) A matrix A is symmetric if it is equal to its transpose AT , obtained by inter-
changing the rows and columns of A. Let Sn be the set of all symmetric n × n matrices. Show that Sn is
a subspace of Mn,n and that dim Sn = n(n + 1)/2.
44. (Sums and intersections) Let H and K be subspaces of a vector space V and define
H + K = {h + k : h ∈ H and k ∈ K}.
(a) Show that H + K is a subspace of V .
(b) Assume that H and K are each finite dimensional, and that H ∩ K = {0}. Show that dim(H + K) =
dim(H) + dim(K).
45. (Hard) (Wronskian) Let f and g be differentiable functions from R to R.
(a) Show that f and g are linearly independent if the determinant

f (x) g(x)
0
f (x) g 0 (x)

is non-zero for some x in R.
(b) Show that the functions sin x and cos x are linearly independent.
(c) Try to generalise the result of (a) to more than two functions.
16
46. (Hard) (Function spaces) Which of the following lie in the complex vector space spanned by f = eix
and g = e−ix , x ∈ R ?
(a) cos x (b) sin x (c) cos x + 3i sin x

(d) sin 2x (e) cosh x
Coordinate vectors, AR 5.4
47. Let V be a vector space and suppose that B is a basis for V . Show that any vector in V can be written
uniquely as a linear combination of elements from B.
48. Find the coordinate vector of v with respect to the given basis B for the vector space V .
(a) v = 2 − x + 3x2 , B = {1, x, x2 , x3 }, V = P3 .

1 2 1
(b) v = , B = {E ij |i = 1, 2; j = 1, 2, 3}, V = M2,3 .
−1 1 2
(c) v = 2 − 5x, B = {x + 1, x − 1}, V = P1

−2 0 1 0 0 0
(d) v = ,B = , , V is the vector space of all diagonal 2 × 2 matrices.
0 3 0 0 0 1
49. (a) Show that the set B = {(−2, 2, 2), (3, −2, 3), (2, −1, 1)} is a basis for R3 .
(b) Find the vectors x, y ∈ R3 whose coordinates with respect to B are

   
2 1
[x]B = 1 and [y]B =  0 .
1 −1
(c) For each of the following vectors find its coordinates with respect to B:
a = (2, −1, 1), b = (1, 0, 5), c = (3, −1, 6).
50. Use coordinate vectors to decide whether or not the given set is linearly independent. If it is linearly
dependent, express one of the vectors as a linear combination of the others.
(a) {x2 + x − 1, x2 − 2x + 3, x2 + 4x − 3} ∈ P2

1 2 0 −1 1 0
(b) , , in M2,2
−1 0 1 1 1 2
17
MAST10018: Linear Algebra Extension Studies Exercises 4: Inner product spaces
Exercises 4: Inner product spaces

Inner product spaces , AR 6.1

x1 y1
1. In R2 , for x = and y = , define hx, yi = x1 y1 + 3x2 y2 . Show that hx, yi is an inner product
x2 y2
on R2 .
2. In R2 , let hx, yi = x1 y1 − x2 y2 . Is this an inner product? If not, why not?
3. Verify that the operation

hx, yi = x1 y1 − x1 y2 − x2 y1 + 3x2 y2
where x = (x1 , x2 ) and y = (y1 , y2 ) is an inner product in R2 .
4. Decide which of the suggested operations on x = (x1 , x2 , x3 ) and y = (y1 , y2 , y3 ) in R3 define an inner
product:
(a) hx, yi = x1 y1 + 2x2 y2 + x3 y3 , (b) hx, yi = x21 y12 + x22 y22 + x23 y32 ,
(c) hx, yi = x1 y1 − x2 y2 + x3 y3 , (d) hx, yi = x1 y1 + x2 y2 .
5. Decide which of the operations hp, qi on real polynomials p(x) = a0 +a1 x+a2 x2 and q(x) = b0 +b1 x+b2 x2
define an inner product:
(a) hp, qi = a0 b0 + a1 b1 + a2 b2 (b) hp, qi = a0 b0

Z 1
(c) hp, qi = p(x)q(x) dx
0

u1 u2 v1 v2
6. (a) If U = and V = are any two 2 × 2 matrices, then
u3 u4 v3 v4
hU, V i = u1 v1 + u2 v2 + u3 v3 + u4 v4
defines an inner product on M2,2 .

3 −2 −1 3
Compute hU, V i if U = and V = .
4 8 1 1
(b) If p = a0 + a1 x + a2 x2 and q = b0 + b1 x + b2 x2 are any two vectors in P2 , then
hp, qi = a0 b0 + a1 b1 + a2 b2
is an inner product on P2 . Compute hp, qi if p = −2 + x + 3x2 and q = 4 − 7x2 .
7. For the vectors x = (1, 1, 0), y = (0, 1, 0) in R3 compute the norms kxk and kyk as well as the angle
between x and y using the following inner products.
(a) hx, yi = x1 y1 + x2 y2 + x3 y3 (b) hx, yi = x1 y1 + 3x2 y2 + x3 y3
8. Let P2 have the inner product defined in Q6b. If p = −2 + 3x + 2x2 , find kpk.

−2 5
9. Let M2,2 have the inner product defined in Q6a. If A = , find kAk.
3 6
10. Let P2 have the inner product defined in Q6b. If p = 3 − x + x2 , q = 2 + 5x2 , find d(p, q).

2 6 −4 7
11. Consider M2,2 with the inner product defined in Q6a. If A = and B = , find
9 4 1 6
d(A, B).
18
12. (Parallelogram law) Prove that the following holds for all vectors x, y in every inner product space.
kx + yk2 + kx − yk2 = 2kxk2 + 2kyk2
Give a geometric interpretation of this result.

[Hint: consider lengths of edges connecting vertices of a parallelogram with sides given by the vectors x
and y.]
13. (Quadratic forms) Let A be a real invertible n × n matrix. Show that
hx, yi ≡ yT AT Ax = (Ay)T (Ax)
defines an inner product in Rn , where x and y are column vectors in Rn . What happens when A is not
invertible? (Note: M T is the transpose of a matrix M , obtained by interchanging the rows and columns
of M .)
14. (Hard) By choosing appropriate vectors in the Cauchy-Schwartz inequality, prove that
(a1 + . . . + an )2 6 n(a21 + . . . + a2n )
for all real numbers a1 , . . . , an . When does equality hold?
Orthogonality and orthonormal bases, AR6.2, AR6.3
15. In each part determine whether the given vectors are orthogonal with respect to the Euclidean inner
product (i.e., the usual dot product).
(a) u = (−1, 3, 2), v = (4, 2, −1) (b) u = (0, 3, −2, 1), v = (5, 2, −1, 0)
16. Consider R2 and R3 each with the Euclidean inner product. In each part find the cosine of the angle
between u and v.
(a) u = (1, −3), v = (2, 4) (b) u = (−1, 5, 2), v = (2, 4, −9)
17. Show that p = 1 − x + 2x2 and q = 2x + x2 are orthogonal with respect to the inner product in Q6b.

2 1
18. Let A = . Which of the following matrices are orthogonal to A with respect to the inner product
−1 3
in Q6a?
   
1 1 2 1
(a)   (b)  
0 −1 5 2
19. Show that in every real inner product space: v + w is orthogonal to v − w if and only if kvk = kwk. Give
a geometric interpretation of this result.
20. Let hx, yi be an inner product on a vector space V , and let e1 , e2 , . . . , en be an orthonormal basis for V .
Prove:
(a) For each x ∈ V , x = hx, e1 ie1 + hx, e2 ie2 + · · · + hx, en ien ;
(b) hα1 e1 + α2 e2 + · · · + αn en , β1 e1 + β2 e2 + · · · + βn en i = α1 β1 + α2 β2 + · · · + αn βn ;
(c) hx, yi = hx, e1 ihy, e1 i + · · · + hx, en ihy, en i.
19
21. Express the given vector as a linear combination of the vectors in the orthonormal basis (with respect to
the dot product)
1 2 2 2 1 2 2 2 1
,− , , − , , , , , .
3 3 3 3 3 3 3 3 3
(a) x = (1, 2, 3) (b) y = (−1, 0, 1)
22. (Hard) (Pythagoras’s

p theorem) Let u1 , u2 , . . . , un be orthogonal vectors in an inner product space V
and let kxk = hx, xi be the norm induced by the inner product on V . Show that
ku1 + u2 + · · · + un k2 = ku1 k2 + ku2 k2 + · · · + kun k2 .
23. Use the Gram-Schmidt procedure to construct orthonormal bases for the subspaces of Rn spanned by the
following sets of vectors (using the dot product):
(a) (1, 0, 1, 0), (2, 1, 1, 1), (1, −1, 1, −1)
(b) (2, 2, −1, 0), (2, 3, 1, −2), (3, 4, 5, −2)
(c) (1, −2, 1, 3, −1), (0, 6, −2, −6, 0), (4, −2, 2, 6, −4)
24. Find the orthogonal projection of (x, y, z) onto the subspace of R3 spanned by the vectors
(a) (1, 2, 2), (−2, 2, −1); (b) (1, 2, −1), (0, −1, 2).
25. (Hard) Let V be a finite dimensional real vector space with inner product h , i, and let W be a subspace
of V . Then the orthogonal complement of W is defined as
W ⊥ = {v ∈ V : hv, wi = 0 for all w ∈ W }.
Prove the following:
(a) W ⊥ is a subspace of V .
(b) W ∩ W ⊥ = {0}.
(c) dim W + dim W ⊥ = dim V .
26. Find the least squares polynomial of the specified degree n for the given data points.
(a) {(0, 0), (1, 0), (2, 1), (3, 3), (4, 5)}, n = 1
(b) {(−2, 2), (−1, 1), (0, −1), (1, 0), (2, 3)}, n = 2
27. A maths lecturer was placed on a rack by his students and stretched to lengths L = 1.7, 2.0 and 2.3 metres
when forces of F = 1, 2 and 4 tonnes were applied. Assuming Hooke’s law L = a + bF , find his normal
length a by least squares.
28. A firm that manufactures widgets finds the daily consumer demand d(x) for widgets as a function of their
price x is as in the following table:
x 1 1.5 2 2.5 3
d(x) 200 180 150 100 25
Using least squares polynomials, approximate the daily consumer demand
(a) By a linear function.
(b) By a quadratic function.
20
29. Let P2 be the vector space of polynomials of degree at most two with the inner product
Z 1
hp, qi = p(x)q(x) dx.
−1
Obtain an orthonormal basis for P2 from the basis {1, x, x2 } using the Gram-Schmidt process.
30. (Hard) (Fourier series) Show that the infinite set

n 1 1 1 o
√ , √ sin nx, √ cos nx : n = 1, 2, . . .
2π π π
is an orthonormal set in the vector space C[0, 2π] of real continuous functions on the interval [0, 2π]
equipped with the inner product Z 2π
hf, gi = f (x)g(x) dx.
0
31. (Hard) Find the least squares approximation of f (x) = 1 + x in the vector space C[0, 2π] equipped with
the inner product of Q30 by
(a) a trigonometric polynomial of order 2 or less.
(b) a trigonometric polynomial of order k or less.
[Suggestion: Use the orthonormal basis of the preceding problem.]

Z Z Z Z
Note that xn sin x dx = −xn cos x + n xn−1 cos x dx and xn cos x dx = xn sin x − n xn−1 sin x dx.
21
MAST10018: Linear Algebra Extension Studies Exercises 5: Linear transformations
Exercises 5: Linear transformations

Linear transformations AR8.1–3
1. Show that each of the following functions is a linear transformation:
(a) S : R2 → R2 , S(x, y) = (2x − y, x + y),

y z
(b) T : R3 → M2,2 given by T (x, y, z) = .
−x 0
2. Determine whether or not the given function is a linear transformation, and justify your answer.
(a) F : R3 → R2 , F (x, y, z) = (0, 2x + y) (b) K : R2 → R3 , K(x, y) = (x, sin y, 2x + y)
3. Let v1 , v2 , and v3 be vectors in a vector space V , and T : V → R3 a linear transformation for which
T (v1 ) = (1, −1, 2), T (v2 ) = (0, 3, 2), and T (v3 ) = (−3, 1, 2). Find T (2v1 − 3v2 + 4v3 ).
4. For the linear transformations of R2 into R2 given by the following matrices:

(i) Sketch the image of the rectangle with vertices (0, 0), (2, 0), (0, 1), (2, 1).
(ii) Describe the geometric effect of the linear transformation.
(note that the constants a, b, c are assumed to be non-zero)
     
0 1 0 0 1 1
(a)   (b)   (c)  
1 0 1 0 0 0
     
1 0 b 0 3 −4
(d)   (e)   (f) 51  
a 1 0 c 4 3
5. (Geometric transformations) Find the matrix of the following linear transformations of R2 .
3π
(a) rotation by 4 (b) rotation by − π2
(c) reflection in the line y = x (d) reflection in the x−axis
6. In each part, find a single matrix that performs the indicated succession of operations.
1
(a) Compresses by a factor of 2 in the x-direction, then expands by a factor of 5 in the y-direction.
(b) Reflects about y = x, then rotates about the origin through an angle of π.
(c) Reflects about the y-axis, then expands by a factor of 5 in the x-direction, and then reflects about
y = x.
The following functions are all linear transformations. Use them in Qs 7 to 9.
K(x, y, z) = (x, x + y, x + y + z) L(x, y, z) = (2x − y, x + 2y)

S(x, y, z) = (z, y, x) T (x, y) = (2x + y, x + y, x − y, x − 2y)
7. Find the matrix that represents each of the following linear transformations (with respect to the standard
bases).
(a) K (b) L (c) S (d) T
8. (Combining linear transformations) Find the indicated linear transformation if it is defined. If it is

not defined, explain why it is not.
22
(a) LK(= L ◦ K) (b) T L(= T ◦ L) (c) S 2 (d) K + S (e) T 2
9. Find the matrices that represent those linear transformations in Q7 that exist. (Use your results from the
previous question.)
10. Determine whether or not v1 = (−2, 0, 0, 2) or v2 = (−2, 2, 2, 0) is in the kernel of the linear transformation
T : R4 → R3 given by T (x) = Ax where
 
1 2 −1 1
A = 1 0 1 1 .
2 −4 6 2
11. Determine whether or not w1 = (1, 3, 1) or w2 = (−1, −1, −2) is in the image of the linear transformation
given in Q10.
12. For the linear transformation given in Q10, find the nullity of T and give a basis for the kernel of T . Is
the transformation injective?
13. For the linear transformation given in Q10, find the rank of T and give a basis for the image of T . Is the
transformation surjective?
14. For the linear transformation T find (i) its standard matrix, (ii) a basis for the kernel and (iii) a basis for
the image.
 
    x  
x x+y  1  x + x − x
1 2 3
(a) T   =  (b) T x2  = 
   
y 3y    2x1 + x2
x3
     
  x + 2y x1 3x1 − x2 − 6x3
x      
(c) T   =  −y  (d) T x2  = −2x1 + x2 + 5x3 
     
y      
x−y x3 3x1 + 3x2 + 6x3
15. Determine whether or not the given linear transformation is invertible. If it is invertible, compute its
inverse.
(a) T : R3 → R3 given by T (x, y, z) = (x + z, x − y + z, y + 2z)
(b) T : R2 → R2 given by T (x, yz) = (3x + 2y, −6x − 4y)
(c) Tθ : R2 → R2 an anticlockwise rotation around the origin through an angle of θ.
(d) T θ : R2 → R2 a reflection in the line through the origin which forms an angle θ with the x-axis.
16. Show that the transformation    

x x+y
T y  =  y + z 
z z+x
is invertible and find its inverse.
23
17. Consider the matrix  

cos θ − sin θ 0
A(θ) =  sin θ cos θ 0 .
0 0 1
(a) Evaluate det A(θ).
(b) Geometrically interpret the effect of multiplying a vector by A(θ).
(c) Show that A(θ)A(φ) = A(θ + φ) and interpret this result.
(d) Use the previous part to find the inverse of A(θ). How does this compare with the transpose A(θ)T ?
18. Find the matrices of the transformations T that orthogonally project a point (x, y, z) onto the following
subspaces of R3 . Show by two methods that each transformation is idempotent (i.e., T ◦ T = T ).
(a) The z-axis.
(b) The straight line x = y = 2z.
(c) The plane x + y + z = 0.
19. (Computer graphics) One of the most important applications of linear transformations is computer
graphics where we wish to view 3-dimensional objects (for example a crystal) on a 2-dimensional screen.
The screen is the xy-plane. The aim is to rotate the crystal and orthogonally project it onto the xy-plane
to obtain different views of it. We consider 3 possible rotations:
 
1 0 0
• A rotation of θ round the x-axis using the matrix Rx = 0 cos θ − sin θ
0 sin θ cos θ
 
cos θ 0 sin θ
• A rotation of θ round the y-axis using the matrix Ry =  0 1 0 
− sin θ 0 cos θ
 
cos θ − sin θ 0
• A rotation of θ round the z-axis using the matrix Rz =  sin θ cos θ 0
0 0 1

1 0 0
The matrix used to orthogonally project the object onto the computer screen is P = .
0 1 0
The crystal has vertices A : (0, 0, 1), B : (1, 0, 0), C : (0, 1, 0), D : (−1, 0, 0), E : (0, −1, 0), F : (0, 0, −1)
with edges the line segments AB, AC, AD, AE, F B, F C, F D, F E, BC, CD, DE and EB.
z
A
D
E C y
B
F
x
 
0
1 0 0   0
Projecting onto the xy-plane, we find P (A) = A0 is given by P x = 0 = ,
0 1 0 0
1
24
so that A0 = (0, 0). In the same way B 0 = (1, 0), C 0 = (0, 1), D0 = (−1, 0), E 0 = (0, −1) and F 0 = (0, 0).
Connecting up the line segments appropriately we find the view on the computer screen is
y
A’& F’
C’
D’ B’ x
E’
Draw the picture that would appear on the computer screen if:
(a) the crystal is rotated 45◦ around the x-axis before projection,
(b) the crystal is rotated 45◦ around the x-axis and then 30◦ around the z-axis,
(c) the crystal is rotated 45◦ around the x-axis, 30◦ around the z-axis and then −60◦ around the y-axis.
You might like to think about whether or not orthogonal projection is the best way to project from 3-D
to 2-D.
Matrix representations, AR8.4
20. Let Pn denote the vector space of all real polynomials in the variable x of degree 6 n. Let T : P2 → P3
denote the function defined by multiplication by x: T (p(x)) = xp(x). In other words, T (a + bx + cx2 ) =
ax + bx2 + cx3 .
(a) Show that T is a linear transformation.
(b) Find the matrix of T with respect to the standard bases {1, x, x2 } for P2 and {1, x, x2 , x3 } for P3 .
21. Let T : P2 → P2 be the linear transformation defined by T (p(x)) = p(2x + 1), that is,
T (a0 + a1 x + a2 x2 ) = a0 + a1 (2x + 1) + a2 (2x + 1)2 .
Find [T ]B with respect to the basis B = {1, x, x2 }.
22. Find the matrix A that represents the linear transformation T with respect to the bases B and B 0 .
(a) T : R3 → M2,2 given by

y z
T (x, y, z) =
−x 0
where B = {e1 , e2 , e3 } and B 0 = {E ij |i = 1, 2; j = 1, 2} (i.e. the standard basis for M2,2 ).
(b) T : P3 → P3 given by
T (a0 + a1 x + a2 x2 + a3 x3 ) = (a0 + a2 ) − (a1 + 2a3 )x2
where B, B 0 = {1, x, x2 , x3 }.
23. Compute ker(T ) for each linear transformation in Q22. Which ones are one-to-one?
24. Compute Im (T ) for each linear transformation in Q22. Which ones are onto?
25. Let V be the vector space of all real 2 × 2 matrices. Let T : V → R2 be the map defined by

a11 a12 a11 a12
T = 2 −1 = 2a11 − a21 2a12 − a22 .
a21 a22 a21 a22
(a) Show that T is a linear transformation.
25
(b) Find bases for the kernel and image of T . Deduce the rank and nullity of T .
(c) Find the matrix of T with respect to the basis

1 0 0 1 0 0 0 0
, , ,
0 0 0 0 1 0 0 1
of V and the standard basis of R2 .
26. Let S : P2 → P3 be defined as follows. For each p(x) = a2 x2 +a1 x+a0 , define S(p) = 31 a2 x3 + 21 a1 x2 +a0 x.
Find the matrix A that represents S with respect to the bases B = {1, x, x2 } and B 0 = {1, x, x2 , x3 } (The
linear transformation S gives the integral of p(x), with the constant term equal to zero.)
27. Use the matrix of Q26 to find the integral of p(x) = 1 − x + 2x2 .
Change of basis, AR8.5,6.5
28. (a) Find the transition matrix P from B to C, where

B = {(1, −2, 1), (0, 3, 2), (1, 0, −1)} and C = {(1, 0, 0), (0, 1, 0), (0, 0, 1)}.
(b) Use P to find [x]B (the coordinate vector of x with respect to B) if
(a) x = (3, −2, 5) (b) x = (−2, 7, 4)
29. Verify that the given set B is a basis for Rn . Compute the change of basis matrix for each of the bases,
and use it to find the coordinate vector of v with respect to B.
(a) B = {(1, 2), (1, −2)}, v = (−1, 3) (n = 2)
(b) B = {(1, 1, 1), (1, 0, 1), (−1, 1, 0)}, v = (3, −1, 1) (n = 3)
30. Let T : Rn → Rn be given by T (x) = Ax where A is the given matrix.

1 1
(a) A =
0 −1
Find the matrix that represents T with respect to the basis B of Q29a.
 
1 1 3
(b) A = 2
 −1 1
1 2 0
Find the matrix that represents T with respect to the basis B of Q29b.
31. Let T : R3 → R3 be given by T (x) = T (x, y, z) = (2x − y, x + y + z, y − z). Find the matrix that represents
T with respect to the basis B of Q29b.
32. Write down the matrix of T with respect to B, and compute the matrix of T with respect to B 0 , where
T : R2 → R2 is defined by T (x1 , x2 ) = (x1 − 2x2 , −x2 ), B = {u1 , u2 }, B 0 = {v1 , v2 }, and u1 = (1, 0),
u2 = (0, 1), v1 = (2, 1), v2 = (−3, 4).
33. A linear transformation T : R3 → R3 has matrix

 
2 3 0
−1 1 2
2 0 1
with respect to the standard basis for R3 . Find the matrix of T with respect to the basis
B = {(1, 2, 1), (0, 1, −1), (2, 3, 2)}.
26
Cayley-Hamilton Theorem
34. Verify the Cayley-Hamilton Theorem for the following matrices.

 
  0 1 0
3 6  
(a)  (b) 0 0 1
  
1 2  
1 −3 3
 
0 1 0
35. Use the Cayley-Hamilton Theorem to calculate the inverse of the matrix 0 0 1 .
1 −3 3
36. For each matrix, find a non-zero polynomial satisfied by the matrix.
 
  1 4 −3
2 5  
(a)  (b) 0 3 1 
  
1 −3  
0 2 −1
27
MAST10018: Linear Algebra Extension Studies Exercises 6: Eigenvalues and eigenvectors
Exercises 6: Eigenvalues and eigenvectors
We reserve u1 , u2 , . . . for eigenvectors, λ1 , λ2 , . . . for eigenvalues, x1 , x2 , . . . for components of vector x; and write
W (λ) for the eigenspace corresponding to λ.
Eigenvectors, eigenvalues and diagonalisation, AR7.1–3, AR9.6–7
1. Find the eigenvalues and linearly independent eigenvectors of the following matrices.
       
2 0 7 −2 1 −1 0 −1
(a)   (b)   (c)   (d)  
−1 2 15 −4 1 3 1 0
2. Find 1-dimensional subspaces of R2 invariant under the operation of the following matrices:
   
1 0 1 2
(a)   (b)  
2 0 −3 −6
3. (Rotations) Show that there is no line in the real plane R2 through the origin which is invariant under
the transformation whose matrix is
cos θ − sin θ
A(θ) = ,
sin θ cos θ
when θ is not an integral multiple of π. Give a geometric interpretation of this problem commenting on
the case when θ = kπ for some k ∈ Z.
4. Find the eigenvalues of the given matrix by inspection.

 
  1 0 0 0
1 2 0 



  2 3 0 0
(a) 0 3 −1 (b) 
  
 
  4 5 6 0
0 0 4  
7 8 9 10
5. Prove that for an invertible matrix A, λ is an eigenvalue of A if and only if 1

λ is an eigenvalue of A−1 .
What relationship holds between the eigenvectors of A and A−1 ?
6. Find the eigenvalues and linearly independent eigenvectors for the following matrices.
   
2 −3 6 2 1 0
   
(a) 0 5 −6 (b) 0 2
   
0
   
0 1 0 0 0 −3
   
−5 −8 −12 2 2 2
   
(c) −6 −10 −12 (d) −1 −1 −2
   
   
6 10 13 1 2 3
7. For each matrix find all eigenvalues and a basis for each eigenspace.
   
3 1 1 1 1 0
   
(a) 2 (b) 0 1 0
   
4 2
   
1 1 3 0 0 1
28
8. Decide which of the matrices A in Qs 1 and 6 above are diagonalizable and, if possible, find an invertible
matrix P and a diagonal matrix D such that P −1 AP = D.
9. (Hard) Let D be a diagonal matrix, D = diag (λ1 , . . . , λs ) (i.e., Dii = λi , Dij = 0 for i 6= j). Prove by
induction that, for each positive integer n,
Dn = diag (λn1 , . . . , λns ).
10. (Hard) Let A be a matrix such that A = P DP −1 , where D is diagonal. Prove by induction that, for each
positive integer n,
An = P Dn P −1 .
11. Use the results of the preceding two problems to find A5 , where A is
 
  8 1 27 5
  9 18 −24  
−2 −6
 
3   
1 
0 18 14 
(a)  (b) 7 20 −24 (c) 8 
   
−2 0 2 −18 −6

2  
7 21 −25  
0 8 8 −8
[Hint: The matrix has eigenvalues 3, 2, −1 in (b), and 1, 2, −1, −2 in (c).]
12. Suppose the nth pass through a manufacturing process is modelled by the linear equations xn = An x0 ,
where x0 is the initial state of the system and

1 3 2
A= .
5 2 3
Show that " # n " 1 #

1 1
n 2 2 1 2 − 21
A = 1 1
+ .
2 2
5 − 12 1
2

p
Then, with the initial state x0 = , calculate limn→∞ xn .
1−p
13. (Hard) (Fibonacci numbers) The Fibonacci sequence 0, 1, 1, 2, 3, 5, 8, 13, . . . is described

by the
difference
Fk+1
equation Fk+2 = Fk+1 + Fk and the initial conditions F0 = 0, F1 = 1. Writing uk = , show that
Fk
uk+1 = Auk where
1 1
A= .
1 0

1
Solve for uk in terms of u0 = and show that
0
√ !k √ !k
1 1+ 5 1− 5
Fk = √ − .
5 2 2
Hence find the limit as k → ∞ of the ratio Fk+1 /Fk .
29
14. Two companies, Lemon and LIME, introduce a new type of computer. At the start, their shares of the
market are 60% and 40%. After a year, Lemon kept 85% of its customers and gained 25% of LIME’s
customers; LIME gained 15% of Lemon’s customers and kept 75% of its customers. Assume that the total
market is constant and that the same fractions shift among the firms every year.
(a) Write down the market share shift as a system of linear equations.
(b) Express the shift in matrix form and find the transition matrix A.
(c) Find the market shares after 5 and 10 years.
(d) Show that the market eventually reaches a steady state, and give the limit market shares.
15. (Hard) Prove by induction that the following statements are true for all positive integers n.
(a) If λ is an eigenvalue of the square matrix A, then λn is an eigenvalue of the matrix An .
(b) (A−1 )n = (An )−1 if A is invertible.
(c) (AT )n = (An )T for any square matrix A.
(d) (AB)n = An B n , provided that AB = BA. (What happens when the matrices A and B do not
commute?)
16. Determine whether or not the given matrix A is orthogonal.

 
1 2 2
 
3 3 3  0 1 0
2  
1 2

(a)  −3 (b) 1
 
3 3 0 1
   
2
3 − 23 1
3
0 1 0

cos θ − sin θ
17. Show that the rotation matrix A = is orthogonal.
sin θ cos θ
18. (Symmetric matrices) For each symmetric matrix A below find a decomposition A = P DP T , where P
is orthogonal and D diagonal.
   
7 2 0 −2 0 −36
   
(a) 2 6 2 (b)  0 −3
   
0
   
0 2 5 −36 0 −23
   
1 1 0 4 2 2
   
(c) 1 1 0 (d) 2 4 2
   
   
0 0 0 2 2 4
19. Let A be an orthogonal matrix. Show that det A = ±1.
20. Prove that if A, B are orthogonal n × n matrices, then so are A−1 and AB.
30
MAST10018: Linear Algebra Extension Studies Answers 1: Linear equations and matrices
Answers 1: Linear equations and matrices

 
1 2
1. (a)  −5 −1 
7 15
(b) Matrices incompatible.

 
0
(c) 7 + i 6 −4 + 3i 1 − i = (7 + i)0 + 6(1 − i) + (−4 + 3i)1 = 2 − 3i
1
 
0 0 0
(d)  7 6 −4 
7 6 −4
      
3 −6 0 2 3 × 2 + (−6) × 1 + 0 × 1 0
(e) 0 2 −2 1 =  0 × 2 + 2 × 1 + (−2) × 1  = 0
1 −1 −1 1 1 × 2 + (−1) × 1 + (−1) × 1 0
 
3
(f)  5 
7
 
2 0 1
2. A−1 = 0 1
3
2
3

−1 0 −1
3. Verify that

cos θ1 − sin θ1 cos θ2 − sin θ2 cos θ1 cos θ2 − sin θ1 sin θ2 − cos θ1 sin θ2 − sin θ1 cos θ2
=
sin θ1 cos θ1 sin θ2 cos θ2 sin θ1 cos θ2 + cos θ1 sin θ2 −
sin θ1 sin θ2 + cos θ1 cos θ2
cos(θ1 + θ2 ) − sin(θ1 + θ2 )
=
sin(θ1 + θ2 ) cos(θ1 + θ2 )
using the standard sum formulae for sin and cos.

   
2 3 1 −1
3 −3
4. A =  3 4  ; B =  −1 1  ; AT B =
4 −4
4 5 1 −1
   
28 0 0 29 0 0
   
5. (a)  0 29 0 (b)  0 29 0
   
 
   
0 0 30 0 0 29
   
1 2 3 1 2 3
   
(c)  2 4 5  (d)  0 4 5 
   
   
3 5 30 0 0 30

0 1
6. (a)
0 0

0 1
(b)
−1 0
31

1 1 1 −1
(c) C = ,D=
1 1 1 −1

1 0 0 1
(d) E = ,F =
0 −1 1 0
7. If B is the first of the two choices, then the matrix equation AB = BA is

a b 1 0 1 0 a b a 0 a b
= that is = .
c d 0 0 0 0 c d c 0 0 0
Equating coefficients, we have b = c = 0. Repeating with the other choice of B, we have

a b 0 1 0 1 a b 0 a c d
= that is = .
c d 0 0 0 0 c d 0 c 0 0
Again equating coefficients, we have c = 0 and a = d. Thus, if AB = BA for all 2 × 2 matrices B (more
particularly for the two given) then A must be of the form

a 0 1 0
=a = aI2 .
0 a 0 1
That is, A is a scalar matrix.
8.
(AB − ABA)2 = (AB)(AB) − (AB)(ABA) − (ABA)(AB) + (ABA)(ABA)

= ABAB − ABABA − ABAAB + ABAABA
= ABAB − ABABA − ABAB + ABABA = 0.
9. (a) Linear (b) Non linear (c) Linear (d) Non linear
10. An RE form and the RRE form are given. Note RE forms are not unique.
   
1 −2 4 1 0 0
(a) REF:  0 −1 2 , RREF:  0 1 0 
0 0 3 0 0 1

4 − 3i 4 − 3i 5 − 10i 8 − 11i 1 1 0 1.6 + 0.7i
(b) REF: , RREF:
0 0 −10 + 10i −3 + 11i 0 0 1 0.7 − 0.4i
   
1 2 1 0
 0 3   0 1 
(c) REF: 
 0
, RREF:  
0   0 0 
0 0 0 0
   
1 0 −3 2 1 0 0 11
(d) REF:  0 2 1 4 , RREF:  0 1 0 12 
0 0 2 6 0 0 1 3
(e) Reducing to REF and then RREF (last step) we have
(1) : r2 → r2 − 2r1
   
1 2 0 1 1 2 0 1
2 4 1 1 ∼ 0 0 1 −1  r3 → r3 − 3r1
(1)
3 6 1 1 0 0 1 −2 (2) : r3 → r3 − r2
   
1 2 0 1 1 2 0 0 (3) : r1 → r1 + r3
∼ 0 0 1 −1 ∼ 0 0 1 0 r2 → r2 − r3
(2) (3)
0 0 0 −1 0 0 0 1 r3 → −r3
32
   
1 −1 1 1 1 −1 0 0
 0 0 6 21   0 0 1 0 
(f) REF: 
 0
, RREF:  
0 0 33   0 0 0 1 
0 0 0 0 0 0 0 0
  
x 0
11. (a) rank 3;  y  =  0 
z 0
    
x −1 −1.6 − 0.7i
 y   1   0 
(b) rank 2;   , a, b ∈ R
 z  = a  0  + b  −0.7 + 0.4i
   

w 0 1

x 0
(c) rank 2; =
y 0

  
x −11
 y   −1 
 z  = t  −3  , t ∈ R
(d) rank 3;   2 
w 1
  
x −2
 y   1 
 z  = t 0 ,t ∈ R
(e) rank 3;   
w 0
   
x 1
 y   1 
(f) rank 3; 
  = t
 ,t ∈ R
z  0 
w 0
12. 1 and 2.
     
x −1 + 2i 1+i
 = t  1 − 2i  + s  3 − 5i  , s, t ∈ C
 y     
13. (a)  z   2   0 
u 0 4
(b) The corresponding coefficient matrix is

 
7 7 −16 6
4 −1 3 7 .
3 1 −2 4
33
Reduce this matrix to RRE form:
− 16 6 r1 → 71 r1
   
7 7 −16 6 1 1 7 7
4 −1 3 7 ∼ 4 −1 3 7 r2 → r2
3 1 −2 4 3 1 −2 4 r3 → r3

1 1 − 16 6
 r1 → r1
7 7
∼ 0

−5 85 25  r2 → r2 − 4r1
7 7 
0 −2 34 10 r3 → r3 − 3r1
7 7

1 1 − 16 6
 r1 → r1
7 7
− 17 − 57  r2 → − 15 r2
∼ 0 1
 
7
0 −2 34 10 r3 → r3
7 7

1 0 1 11
 r1 → r1 − r2
7 7
r2 → r2
∼ 0 1 − 17 − 75 
 
7
0 0 0 0 r3 → r3 + 2r2
So
1 11
x + 7z + 7 w =0
17 5
y− 7 z − 7w =0
The unknowns z and w (corresponding to the columns without leading entries) can be chosen as
parameters:
z = 7α, w = 7β;
the coefficient 7 is introduced to avoid fractions. Then
1 11 17 5
x = − z − w = −α − 11β, y= z + w = 17α + 5β,
7 7 7 7
or,      
x −1 −11
y 
  = α  17  + β  5 
   
z  7  0 
w 0 7
for arbitrary real scalars α, β.
14. (a) and (b) rank(A|B) > rank A: no solution, so the system consists of three planes that do not all
intersect in any points.
     
x 23 −20
(c) rank(A|B) = rank A = 2: multiple solutions: y  = α  35 2
 + − 31 ;
2
z 1 0
The system consists of three planes intersecting in a line.
(d) We reduce the augmented matrix of the system of equations by elementary row operations that avoid
fractions:
   
2 −3 5 10 2 −3 5 10
 4 7 −2 −5  ∼  0 13 −12 −25  r2 → r2 − 2r1
r3 → r3 − r1
2 −4 25 31 0 −1 20 21
 
2 −3 5 10
∼ 0 −1 20 21  r2 ↔ r3
0 13 −12 −25
 
2 −3 5 10
∼ 0 −1 20 21  r3 → r3 + 13r2
0 0 248 248
34
So rank A = 3 = number of variables and the number of parameters required for the general solution is
d = 3 − rank A = 0. Thus the equations have a unique solution. Continuing to RRE form we get
 
1 0 0 1
 0 1 0 −1  .
0 0 1 1
Thus    
x 1
 y  = −1
z 1
is the unique solution; the system consists of three planes intersecting in a single point.
       
x1 −2 −2 −3
x2  1 3 9
x3  = α  1  + β  0  +  0  , α, β ∈ R
15.        
x4 0 1 0
           
x 1 x −1 −1 1
1
16. (a) (i) k = −2: no solution; (ii) k ∈ R\{−2, 1}: y  = 1; (iii) k = 1: y  = α  1  + β  0  + 0.
k+2
z 1 z 0 1 0
         
x k−2 x 3 3
1
(b) (i) k = 2: no solution; (ii) k ∈ R\{−1, 2}: y  =  1 ; (iii) k = −1: y  = α 2 + 1.
k−2
z 2 z 1 0
(c) In reducing the augmented matrix we swap the first and third rows initially. This is so the expressions
in k do not become too unwieldy.
     
2 k−4 3−k 1 1 −2 1 1 1 −2 1 1
 0 2 k−3 2  ∼  0 2 k−3 2  ∼  0 2 k−3 2 
r1 ↔r3 r3 →r3 −2r1
1 −2 1 1 2 k−4 3−k 1 0 k 1 − k −1
   
1 −2 1 1 1 −2 1 1
∼  0 2 k−3 2 = 0 2 k−3 2 
r1 ↔r3 2
0 0 k − k − 2 2k + 2 0 0 (k − 2)(k + 1) 2(k + 1)
(i) The system has no solution when there is a row of zeroes in the coefficient matrix corresponding
to a non-zero entry in the augmented component (we may think of this as the rank of the augmented
matrix being greater than the rank of the coefficient matrix). This occurs when
(k − 2)(k + 1) = 0 and 2(k + 1) 6= 0
i.e. for k = 2.
(ii) The system has a unique solution when there is no zero row matched to a non-zero augmented
component (ie. the rank of the coefficient matrix and the rank of the augmented matrix are equal)
and the rank is equal to the number of unknowns. This occurs for (k − 2)(k + 1) 6= 0, i.e. for
k ∈ R\{−1, 2}. In this case we are free to divide the last row through by (k − 2)(k + 1), and then to
proceed to RRE form.
 
1 0 k−2 3 r1 → r1 + r2
r2
∼  0 1 k−3 2 1  r2 → 2
2 r3
0 0 1 k−2
r3 → (k+1)(k−2)
 
1 0 0 1
1 r1 → r1 − (k − 2)r3
∼ 0 1 0
r2 − k−3
k−2

2 r2 → 2 r3
0 0 1 k−2
35
So for k ∈ R\{−1, 2}    
x k−2
y  = 1  1 
k−2
z 2
is the unique solution. Note that a unique solution does not mean that there is only “one” answer -
in this case it means that for each k value k ∈ R\{−1, 2} there is a unique solution.
(iii) The system has an infinite number of solutions if we again do not have a row of zeros with a
non-zero augmented component, but now the rank is less than n, the number of unknowns. The
number of parameters required for the general solution is equal to d = n − rank A. This occurs for
k = −1. (The third row of the RE form becomes a row of zeros). Substituting k = −1 in the RE
form and reducing to RRE form we have
 
1 0 −3 3
∼  0 1 −2 1 
0 0 0 0
Then rank A = 2, and we need d = 3 − 2 = 1 parameters for the general solution. Choosing the
variable z (which corresponds to the column without a leading entry) as the free parameter λ, and
rearranging we have
x = 3 + 3λ, y = 1 + 2λ, z = λ
or equivalently      
x 3 3
y  = 1 + λ 2 , λ∈R
z 0 1
(occurring for k = −1.)
         
x −3k − 3 x 3 −3
1 ; (iii) k = 2: y  = α − 5  +  2 .
(d) (i) k = −5: no solution; (ii) k ∈ R\{−5, 2}: y  =  4 2
k+5
z 4 z 1 0
     
x −1 a + 2b
1
17. (a) If c = −2a + b: y  = α  11  + 3a − b , α∈R
7
z 7 0
(b) We reduce the augmented matrix [A|B] of the system to RRE form. The system is solvable if and
only if rank A = rank[A|B].
r1 → r1
   
1 −2 4 a 1 −2 4 a
 2 3 −1 b ∼ 0 7 −9 b − 2a  r2 → r2 − 2r1
3 1 2 c 0 7 −10 c − 3a r3 → r3 − 3r1
r1 → 7r1 + 2r2
 
7 0 10 3a + 2b
∼ 0 7 −9 b − 2a  r2 → r2
0 0 1 a+b−c r3 → −r3 + r2
r1 → r1 − 10r3
 
7 0 0 −7a − 8b + 10c
∼ 0 7 0 7a + 10b − 9c  r2 → r2 + 9r3
0 0 1 a+b−c r3 → r3
Therefore rank A = 3 = rank[A|B], and the system is consistent for any choice of constants a, b, c.
For any triple a, b, c, the solution depends on d parameters, where d = n − rank A = 3 − 3 = 0. Hence
there is a unique solution
      
x −7a − 8b + 10c −7 −8 10 a
1 1
y  =  7a + 10b − 9c  =  7 10 −9  b  .
7 7
z 7a + 7b − 7c 7 7 −7 c
36
    
x −7 −8 10 a
1
(c) For all a, b, c: y  =  7 10 −9  b 
7
z 7 7 −7 c
 
0 0 1
18. (a) 0 1 0
1 0 0
 
1 0 0
(b) 0 7 −3
0 −9 4
(c) We augment the matrix by the identity and reduce to RRE form (which is of form [I|A−1 ] if the
inverse exists)
   
4 7 2 | 1 0 0 2 4 1 | 0 0 1
(1) : r1 ↔ r3
[A|I] = −3 1 −7 | 0 1 0 ∼ −6 2 −14 | 0 2 0
(1) r2 → 2r2
2 4 1 | 0 0 1 4 7 2 | 1 0 0
   (2) : r2 → r2 + 3r1
2 4 1 | 0 0 1 2 0 1 | 4 0 −7 r3 → r3 − 2r1
∼ 0 14 −11 | 0 2 3  ∼ 0 1 0 | −1 0 2 
(2)
0 −1 0 | 1 0 −2
(3)
0 0 −11 | 14 2 −25 (3) : r1 → r1 + 4r3
58 2 −102 −51
r2 → −r3
1 0 0 | 29 1
   
2 0 0 | 11 11 11 11 11 11 r3 → r2 + 14r3
∼ 0 1 0 | −1 0 2 ∼ 0 1 0 | −1 0 2  (4) : r 1
→ − 11 r3
 
(4) −14 −2 25 (5) −14 −2 25 3
0 0 1 | 11 11 11 0 0 1 | 11 11 11 r1 1
→ r1 + 11 r3
= [I|A−1 ] (5) : r1 1
→ 2 r1
So the inverse is  
29 1 −51
1 
−11 0 22 
11
−14 −2 25
(d) no inverse

1 3−i 1 − 2i
19. (a)
5 −2 − i 1 + 3i
(b) We repeat the procedure of the previous exercise (omitting the descriptions of the row operations).
−i −i 1
| 21 0

2 −i | 1 0 1 1 | 0
∼ 2 ∼ 2 2
3+i 1 | 0 1 3+i 1 | 0 1 0 12 + 3i2 | − 32 − 2i 1
−i 1 1−3i 3+i
1 2 | 2 0 1 0 | 10 10
∼ −6+8i 2(1−3i ∼ −6+8i 2(1−3i .
0 1 | 10 10 0 1 | 10 10
Thus the inverse is

1 1 − 3i 3+i
.
10 −6 + 8i 2 − 6i
  
17 −11 −13 14 8
1 −1 −2 14 −22 1−9

20.  ;
45  10 −25 −5 −5  5  0 
15 −15 −30 15 10
21. (a) −5 (b) −9 (c) 0 (d) 1
37
(e) Reduction to triangular form:

1 −1 2 3

1
−1 2 3
r1 → r1

0 3 −2 0 = 0
3 −2 0
r1 → r2 − 2r1

0 1 0 0

0
1 0 0
r3 → r3 − r1
0 0 4 1 0 0 4 1 r4 → r4 + 2r1

1 −1 2 3

0 1 0 0
= − r2 ↔ r3 – change of sign
0 3 −2 0

0 0 4 1

1 −1 2 3

0 1 0 0
= − r3 → r3 − 3r2
0 0 −2 0

0 0 4 1

1 −1 2 3

0 1 0 0
= −
0 0 −2 0

0 0 0 1 r4 → r4 + 2r3
= −(1 · 1 · (−2) · 1) = 2
The value of the determinant is then 2.
(f) -1440
22. (a) −1 (b) −1 (c) 3 (d) 8 (e) 1
23. (a) 11 (b) 2 (c) 0

The three vectors in (c) must be coplanar.
24. Proof required.
25. (a) Using determinant operations we factorize

1 a bc 1
a bc
r2 → r2 − r1
1 b ca = 0 b − a ca − bc r3 → r3 − r1

1 c ab 0 c − a ab − bc

1 a bc

= (b − a)(c − a) 0 1 −c
(Taking out common
0 1 −b factors of r2 and r3 )

1 a bc

= (b − a)(c − a) 0 1 −c
r3 → r3 − r2
0 0 −b + c
= (b − a)(c − a)(c − b)
(b) (b − a)(c − a)(c − b)(a + b + c)
26. D2 = λ2 − λ1 , D3 = (λ2 − λ1 )(λ3 − λ1 )(λ3 − λ2 ) and
D4 = (λ2 − λ1 )(λ3 − λ1 )(λ4 − λ1 )(λ3 − λ2 )(λ4 − λ2 )(λ4 − λ3 ).
27. Requires proof. Hint: Take determinants of the relation P 2 = P .
38
MAST10018: Linear Algebra Extension Studies Answers 2: Vectors and geometry
Answers 2: Vectors and geometry

√ √
1. (a) (1, 3, 0); (−22, 42, 9); −25; 41; 110.
√ √
(b) (4, −2, 16); (11, −28, 125); −110; 466; 2 179.
√ √
(c) (1, 0, 1); (5, 9, −4); −1; 2; 6.
2. (a) 0 (b) 6 (c) 5 − 2π

π π √1
3. (a) 2 (b) 3 (c) arccos 15
4. For a ⊥ b we need to find x such that a · b = 0. So 8 − 2x − 2 = 0, i.e. x = 3 makes a and b orthogonal.

For a k b we need a = λb for some λ ∈ R. Considering first and third components of this vector equation
we find 2 = 4λ and 1 = −2λ which has no simultaneous solution for λ. So there is no x such that a and
b are parallel.
5. (a) 5e1 − e2 + 3e3 (b) (10, −13, −7)
6. (a) ± √130 (1, −5, −2)
(b) A vector orthogonal to the given vectors is

(2, 5, 3) × (−1, −5, 3) = (30, −9, −5).
The length of this is p √
302 + (−9)2 + (−5)2 = 1006
Hence the required unit vector is
1
√ (30, −9, −5).
1006
(c) ± √15 (0, 1, −2)
7. Let the points be A, B, C, respectively. The area of the triangle ABC is half the area of the parallelogram
−−→ −→
with sides AB and AC. So
1 −−→ −→ −−→ −→
A = k(OB − OA) × (OC − OA)k.
2
We find the vector
−−→ −→ −−→ −→
w = (OB − OA) × (OC − OA) = (−3, 2, −1) × (0, 3, 1) = 5e1 + 3e2 − 9e3 .
√
Then the area is equal to A = 12 kwk = 12 k(5, 3, −9)k = 12 115. There are two unit vectors orthogonal to
the plane ABC:

w 5 3 −9 −w −5 −3 9
w1 = = √ ,√ ,√ , w2 = = √ ,√ ,√ .
kwk 115 115 115 kwk 115 115 115
8. 31; positively oriented.
9. Take the dot product of the equation

x + (x · b)a = b
with b and expand to obtain x · b = b · b as a · b = 0. Substituting this back into the previous equation
we obtain
x + (b · b)a = b and so x = b − (kbk2 )a.
39
10. Proof required.
11. Proof required.
12. Use the first equality of Q10 with a replaced by a × b, b replaced by c, and c replaced by d. Then we
have
(a × b) · (c × d) = d · (a × b) × c = d · (a.c)b − (a.b)c = (a.c)(b.d) − (a.b)(c.d).
We have used the result of Q11 in the penultimate inequality.
13. (a) Proof required.
(b) By Q12,
(a × b) · (c × d) = (a · c)(b · d) − (a · d)(b · c).
Apply this to each of (b × c) · (a × d), (c × a) · (b × d) and (a × b · (c × d). We get
(b × c) · (a × d) + (c × a) · (b × d)
+(a × b) · (c × d) = (b · a)(c · d) − (b · d)(c · a) + (c · b)(a · d)
−(c · d)(a · b) + (a · c)(b · d) − (a · d)(b · c)
= 0
14. Proof required.
15. (a) The projection of the vector x on the vector a is given by (1/||a||)(x.a)a. In this case, it is
1 1 1
√ (3, −4, 5) · (−3, 1, 2) (−3, 1, 2) = √ (−3)(−3, 1, 2) = √ (9, −3, −6).
14 14 14
(b) ( 15 15
2 , − 2 , 15)
16. (a) Two vectors u and v parallel to the plane are given by
u = (2, 0, 1) − (1, 1, 3) = (1, −1, −2), v = (1, 1, 3) − (−5, 2, 3) = (6, −1, 0)
So w = u × v = (−2, −12, 5) is a normal for the plane. Thus a cartesian equation for the plane has form
−2x − 12y + 5z = d
Substituting a point on the plane, e.g. (2, 0, 1) we find that d = 1. Rewriting we have
2x + 12y − 5z + 1 = 0.
(b) The equation has form x + 2y + 3z = d as (1, 2, 3) is a normal vector. Substituting (2, 0, 1) we find
that d = 5. Rewriting we have: x + 2y + 3z − 5 = 0.
17. 1
18. The angle between the planes is the same as the angle between the normals to the planes. The normals
are given by the two vectors
(2, 1, 3) and (3, −2, 4).
Thus the angle between these two planes is
!
(2, 1, 3) · (3, −2, 4) 16 16
arccos √ p = arccos √ √ = arccos √
22 + 12 + 32 · 32 + (−2)2 + 42 14 · 29 406
40
19. (a) X = (3, 4, −7) + t (1, 2, −3) (b) X = (2, 0, 1) + t (1, −4, 0)
π
20. (a) 3
(b) The directions of the two lines are determined by the vectors (3, 4, 3) and (1, −1, 1) respectively. The
angle, θ, between them is therefore given by:
|(3, 4, 3) · (1, −1, 1)|
cos θ =
k(3, 4, 3)k k(1, −1, 1)k
2 2
= √ √ =√ .
34 3 102
2
So the angle between the lines is given by arccos √102 . Note the absolute value is taken so as to obtain
the acute angle.
21. (a) 3
(b) The line is given by ` = {P ; u}, where u = Q − P = (1, 1, 1). The distance between the line ` and the
point A is then
r
k(A − P ) × uk k(2, −1, 1) × (1, 1, 1)k k(−2, −1, 3)k 14
distance(`, A) = = = √ = .
kuk k(1, 1, 1)k 3 3
22. (a) Writing x − 3 = 4y − 2 = z + 1 = s, we get

1 1
`1 : (x, y, z) = 3, , −1 + s 1, , 1 = A + su.
2 4
Similarly,
`2 : (x, y, z) = (3, 0, 1) + t(−2, 1, 1) = B + tv.
We need a vector orthogonal to both `1 and `2 :
1
u×v = (4, 1, 4) × (−2, 1, 1)
4
3
= − (1, 4, −2)
4
We can use w = (1, 4, −2) to calculate the distance:
−−→
|AB · w| |(0, − 12 , 2) · (1, 4, −2)| 6
distance(`1 , `2 ) = = √ =√ .
kwk 1 + 16 + 4 21
(b) √9
14
23. x = t, y = −5 − t, z = 5 + 3t.
24. The given planes have normal vectors u = (1, 1, 1) and v = (1, −2, 0), respectively, so the line ` of
intersection of the two planes is parallel to the vector w = u × v = (2, 1, −3). We need a point on `;
choose y = 0 and substitute this into the cartesian equations of the planes: x + z = 0, x − 3 = 0. This
gives a point A(3, 0, −3) ∈ `.
Calculate the distance from P (1, 1, 1) to `:

−→ √
kAP × wk k(−2, 1, 4) × (2, 1, −3)k k(−7, 2, −4)k 69
distance(P, `) = = = √ =√ .
kwk k(2, 1, −3)k 14 14

25. arcsin √ 38√ = arcsin √ 38
181 29 5249
41
26. Point of intersection: (6, 4, −12).

−−→
27. Let A(2, −1, 0), B(−5, −3, 1), P (3, 5, −1) and Q(0, 3, −2). The vectors u = AB = (−7, −2, 1) and v =
−−→
P Q = (−3, −2, −1) are parallel to the plane; so
u × v = (7, 2, −1) × (3, 2, 1) = (4, −10, 8) = 2 (2, −5, 4)
is a vector normal to the plane. An equation of the plane is then 2x − 5y + 4z + d = 0; substituting the
coordinates of the point A, we get d = −9. So the plane has a cartesian equation
2x − 5y + 4z − 9 = 0.
The line through the origin orthogonal to the plane has a parametric equation (x, y, z) = t (2, −5, 4).
Substitute this into the cartesian equation of the plane to get the point of intersection:
2 · 2t − 5 · (−5t) + 4 · 4t − 9 = 0
45t − 9 = 0
1
t =
5
So the point of intersection of the line and the plane is 51 (2, −5, 4) = ( 25 , −1, 45 ).
28. Pick a point B in π. For this choose, say, x = 0, y = 0, and calculate z = 1; this gives B(0, 0, 1). Form the
−−→ −→ −−→
vector AB = B − A = (2, −3, 0). The vector AP is then the orthogonal projection of AB in the direction
of the normal vector w = (2, −1, 1) of π:
−−→
−→ −−→ w · AB
AP = Projw AB = w
kwk2
(2, −1, 1) · (2, −3, 0) 7
= w = (2, −1, 1)
6 6
7 7 7
= ,− ,
3 6 6
So
7√

−→ 7 7 7
distance(A, π) = kAP k = (2, −1, 1)
= k(2, −1, 1)k = 6= √ .
6 6 6 6
To get the point P :

−→ 7 7 7 1 11 13
P = A + (P − A) = A + AP = (−2, 3, 1) + ,− , = , , .
3 6 6 3 6 6
Another method:
|ax0 + by0 + cz0 + d| |2 · (−2) − 1 · 3 + 1 · 1 − 1| 7
distance(A, π) = = √ =√ .
k(a, b, c)k 4+1+1 6
29. (a) x − 2y − z + 3 = 0 (b) x + y − z − 6 = 0
30. Proof required.
ab
31. √
a2 + 2b2
32. Suppose that the vertices of the triangle are A, B and C, with position vectors a, b and c. The midpoint
of BC then has position vector (1/2)(b + c) and the line joining A to this midpoint consists of all points
of the form
1
ta + (1 − t) (b + c) .
2
42
Looking for something symmetric with respect to the vertices, we see that
1
(a + b + c)
3
lies on this line (take t = 1/3).
When we repeat this argument replacing A by B and C in turn, it will still be true that (1/3)(a + b + c)
will still lie on the corresponding median. Thus this point lies on all three medians and so they certainly
intersect in a point.
33. Proof required.
34. Proof required.
35. Proof required.
36. There is no loss in generality (and some benefit to our solution) in assuming that A is the origin. Assume
also that B and C have position vectors b and c.
Suppose that L has position vector xb and N has position vector yc. (Recall that they lie on the lines
through the origin parallel to b and c, respectively). Since M lies on the line BC, it must have a position
vector of the form
sb + (1 − s)c.
Since it also lies on the line LN , it must have a position vector of the form
t(xb) + (1 − t)(yc).
Equating these two expressions and observing that b and c are linearly independent, we see that
s = xt and 1 − s = (1 − t)y.
Solving for s, we obtain

x(y − 1)
s= .
y−x
We can now write down the given ratios as
AL x CN 1−y BM 1−s y(1 − x)

= ; = ; = = .
LB 1−x NA y MC s x(y − 1)
The result is now a simple calculation.
43
MAST10018: Linear Algebra Extension Studies Answers 3: Vector spaces
Answers 3: Vector spaces

Linear combinations and linear dependence
1. (a) w1 = 3u + 2v (b) Proof required. (c) c = −2
2. (a) Dependent: v2 = 2v1 (b) Dependent: v4 = −v1 + 3v3

(c) Independent (d) Independent
3. (a) Decide whether (2, −3, 1, −5), (0, 1, 2, 2), (1, −2, 3, 0) are linearly independent: equivalently, show that
the only scalar solution k1 , k2 , k3 ∈ R to the equation
k1 (2, −3, 1, −5) + k2 (0, 1, 2, 2) + k3 (1, −2, 3, 0) = (0, 0, 0, 0)
is the trivial solution k1 = k2 = k3 = 0. In matrix form,

   
2 0 1   0
−3 1 −2 1 k 0
 1 2 3  k2 = 0 .
    
k3
−5 2 0 0
Moving the third row to the top, and applying Gaussian elimination, we obtain
     
1 2 3 1 2 3 1 2 3
−3 1 −2 0 1 1  0 1 1
 2 0 1  ∼ 0 4 5  ∼ 0 0 1 .
     
−5 2 0 0 12 15 0 0 0
The corresponding system has only the trivial solution. Hence the vectors are independent.
(b) Dependent.
4. (a) Dependent (b) Independent

(c) Dependent (d) Independent
5. (a) Dependent (b) Independent (c) Independent
6. Proof required.
Subspaces of Rn
7. (a) Yes, No, No (b) Yes, Yes, Yes (c) No, No, No (d) No, Yes, No.
8. (a) Yes (b) No (c) Yes
(d) The set W is not a subspace of R3 , as it is not closed under multiplication by scalars: (1, 1, 1) ∈ W ,
but (−1).(1, 1, 1) = (−1, −1, −1) ∈
/ W.
(e) W is a subset of the vector space R3 . The closures can be checked (i) directly or (ii) expressing the
generic point of W as a linear combination:
(i) α(a − b, a + b, 2a) + β(c − d, c + d, 2c) = (u − v, u + v, 2u) ∈ W , where u = αa + βc, v = αb + βd.
(ii) Note that W consists of all vectors (a − b, a + b, 2a) = a(1, 1, 2) + b(−1, 1, 0), where a, b ∈ R. Then
W is a vector subspace of R3 by the linear combination theorem.
9. Proofs required.
44
10. Counter-examples required.
Spanning sets and bases
11. (a) Yes (b) No
12. (a) Does not span (b) Does span
13. (The spanning sets obtained below are not the only possible solutions.)
(a) {(2, 0, 0), (0, 1, 0)} (b) {(1, 0, 0), (0, −1, 0), (1, 1, 3)} (c) {(4, 1, 0), (0, 2, −1), (0, 0, 1), (1, 0, 0)}
14. (The spanning sets obtained below are not the only possible solutions.)
(a) {(1, 0, 1, 0), (0, 1, 0, 1)} (b) {(1, 0, 1, 1), (0, 1, −1, 1)}
(c) {(−1, 1, 0), (−1, 0, 1)}
15. (a) Not a basis (b) Is a basis

(c) Is a basis (d) Not a basis (or even a subset)
16. (a) Yes (b) No
17. (The bases given are not the only ones possible.)
(a) {(3, 0, 1), (0, 1, −2)}, dim W = 2
(b) {(1, 0), 90, 10}, dim W = 2
(c) Since the problem asks us to find an arbitrary basis of W , we form a matrix A whose rows are the
given vectors, and reduce it to a RE form matrix B. Then both matrices have the same row space, and
the nonzero rows of B provide a basis for W :
     
−1 2 0 4 1 −2 0 −4 1 −2 0 −4
 3 1 −1 2 0 7 −1 14  0 7 −1 14 
A= −5 3 1 6 ∼ 0 −7 1 −14 ∼ 0 0
     = B.
0 0
7 0 −2 0 0 14 −2 28 0 0 0 0
So dim W = 2, and a possible basis for W is
{(1, −2, 0, −4), (0, 7, −1, 14)}.
18. (a) {u1 , u2 , u3 }; u4 = 2u1 + u2
(b) {u1 , u2 , u4 }; u3 = 2u1 − 3u2
(c) Form the matrix whose columns are the given vectors and reduce to RE form. If we go all the way
to the RRE form, we can determine the coefficients of the linear combination expressing dependent
columns in terms of the linearly independent ones:
   
1 0 2 0 1 0 2 0
 2 3 1 0 ∼ 0 1 −1 0 .
−1 4 −6 2 0 0 0 1
If the original vectors are u1 , u2 , u3 , u4 , in that order, then a basis of their span, selected from them,
is {u1 , u2 , u4 }; moreover, u3 = 2u1 − u2 .
19. (a) Have 4 vectors in a 3-dimensional space.
(b) Two vectors cannot span a 3-dimensional subspace.
45
(c) Four independent vectors span a 4-dimensional space.
(d) Have two linearly independent vectors for a 2-dimensional space.
20. (a) {(1, 1, 1, 0), (−2, −1, 0, 1)}; 2
(b) {(−1, 0, 1, 0, 0), (3, 1, 0, 1, 0), (−2, −1, 0, 0, 1)}; 3
(c) {(1, 0, 1, 0), (0, −1, −2, 3)}; 2
Row, column and solution spaces
21. The dimension of the column space is the number of columns in the RRE form of the matrix which have
leading entries. The dimension of the row space is the number of non-zero rows in the RRE form for the
matrix. Clearly these are the same as having a leading entry in a column means that the row it is in will
be non-empty, and the row reduction method means that all leading entries will be in different rows.
(a) Rank = 2 (b) Rank = 1 (c) Rank = 3
22. (a) {(1, 2), (2, 1} (b) {(1, −1)}

(c) {(1, 0, 1, 2), (−1, 1, 1, −1), (3, 1, 0, 1)}
23. (a) {(1, 2, 1), (2, 1, −1)} (b) {(1, 0, −1)}

(c) {(1, −1, 3), (0, 1, 1), (1, 1, 0)}
24. (a) {(1, −1, 1)} (b) {(1, 0, 1), (0, 1, 0)}
(c) basis = ∅ (= empty set)
      


 0 1   2 
 
       
 
25. (a)  1  ,  0  (b)  3 
     

       
 
 −1

2 
  4 
 
General vector spaces
26. Proof required.
27. (a) Start with x + 0 = x, multiply both sides by λ.
(b) Start with λ + 0 = λ, multiply both sides by x.
(c) Start with 1 + (−1) = 0, multiply both sides by x.
28. (a) Yes (b) Yes (c) No
(d) Let W be the set of real polynomials p (of any degree) with p(0) = 0. Then W is a subset of a known
real vector space V = F(R).
Given any vectors p, q ∈ W and any scalars λ, µ, we form the function
f = λp + µq.
We have to check whether (i) f is a polynomial, and (ii) f (0) = 0.
46
Pn Pk
(i) Let p(t) = i=0 ai ti , and q(t) = i=0 bi t
i
. Take the maximum of n, k, say n, and put
bi = 0 if k < i 6 n.
Pn Pn Pn
Then f (t) = λ i=0 a i ti + µ i
i=0 bi t =
i
i=0 (λai + µbi )t ; so f is a polynomial.
(ii) f (0) = (λp + µq)(0) = λp(0) + µq(0) = λ0 + µ0 = 0.
By the subspace theorem, W is a vector subspace of F(R).
(e) Let W be the set of real polynomials p (of any degree) with p(0) = 1. From the solution to the previous
problem we know that a linear combination f = λp + µq of any two polynomials p, q is again a polynomial.
We check whether f also satisfies the condition f (0) = 1:
f (0) = (λp + µq)(0) = λp(0) + µq(0) = λ · 1 + µ · 1 = λ + µ.
For the choice λ = 2 and µ = 5 the condition is not satisfied, so the vector f = 2p + 5q does not belong
to W . This means that W is not a vector space.
(Note. One counterexample is enough to show that the closure is violated.)
(f) Yes
(g) W is a subset of the vector space F(R) of all functions y : R → R. To check the closures: Let y, z ∈ W ;
then y, z are differentiable functions of the real variable which satisfy the given differential equation. If
u = αy + βz, then
u00 − 3u0 + 2u = (αy + βz)00 − 3(αy + βz)0 + 2(αy + βz)

= αy 00 + βz 00 − 3αy 0 − 3βz 0 + 2αy + 2βz
= α(y 00 − 3y 0 + 2y) + β(z 00 − 3z 0 + 2z) = α · 0 + β · 0 = 0.
This shows that u = αy + βz ∈ W for all y, z ∈ W and all scalars α, β. So W is a vector space by the
vector space theorem.
29. (a) Vector space (b) Not a vector space (No zero vector, not
closed under scalar multiplication, etc.)
(c) Vector space
30. (a) Subspace (b) Not a subspace
31. (a) Subspace (b) Subspace
32. Proof required.
Linear dependence, spanning sets, bases
33. Yes
34. −2p1 + p2 − 2p3
35. (a) Independent (b) 1 + x + 2x2 = (1)(1 + x2 ) + (1)(x + x2 )

(c) Independent (d) 1 = (1) sin2 x + (1) cos2 x
(e) Independent
36. (a) Spans (b) Does not span (c) Spans

(d) Does not span (e) Does not span
47
37. (a) Basis (b) Not a basis

(c) Basis (d) Dependent, does not span
(e) Does not span
38. (a) 2 (b) 2 (c) 3
39. (a) No, not enough vectors as dim(P2 ) = 3 (b) Yes, dim(M2,2 ) = 4 and this set has 4
linearly independent vectors
40. (a) No, correct number of vectors but they are (b) Yes
linearly dependent
41. You can put the vectors as columns in a matrix and then row reduce. If the rank = 4 = dim(M2,2 ) then
the set will span M2,2 and as there are four vectors (just the right number) they will also be linearly
independent if rank = 4
42. (a) Any 4 vectors in a 3-dimensional space are linearly dependent.
(b) No 2 vectors can span a 3-dimensional space.
(c) Let {v1 , v2 , v3 } be a basis for V and {w1 , w2 , w3 } a basis for W . Now any 6 vectors in the 5-dimensional
space R5 must be linearly dependent. Hence there are scalars α1 , α2 , α3 , β1 , β2 , β3 not all zero such that
α1 v1 + α2 v2 + α3 v3 + β1 w1 + β2 w2 + β3 w3 = 0.
Rearranging this equation gives:
α1 v1 + α2 v2 + α3 v3 = −(β1 w1 + β2 w2 + β3 w3 ).
Let u = α1 v1 + α2 v2 + α3 v3 then we have u ∈ V ∩ W . We know that at least one of the coefficients αi , βj

is non-zero, say αi 6= 0. Now the vectors v1 , v2 , v3 form a basis for V so are linearly independent. Hence
we have u = α1 v1 + α2 v2 + α3 v3 6= 0. If some βj 6= 0, a similar argument again shows that u is non-zero.
This proves the result.
43. Proof required.
(b) Hint: Let v1 , . . . , vh be a basis for H and let w1 , . . . , wk be a basis for K. Show that v1 , . . . , vh , w1 , . . . , wk
is a basis for H + K.
45. (a) Hint: Assume that there are scalars α, β ∈ R such that αf + βg is the zero function. This means that
αf (x) + βg(x) = 0 for all x.
Differentiating this gives

αf 0 (x) + βg 0 (x) = 0 for all x.
This gives a system of two linear equations for α, β.
(b) Proof required
(c) Proof required
46. (a) Yes (b) Yes (c) Yes (d) No (e) No
48
Coordinate vectors
47. Proof required.
48. (a) (2, −1, 0, 3) (b) (1, 2, 1, −1, 1, 2)

(c) (− 32 , − 72 ) (d) (−2, 3)
49. (a) To establish that the set B is a basis for R3 , we must show that (i) B is a spanning set for R3 , and
that (ii) B is a linearly independent set.
Write the given vectors a1 , a2 , a3 as columns of a matrix

 
−2 3 2
A = [a1 |a2 |a3 ] =  2 −2 −1 ,
2 3 1
and reduce the matrix to RE form. For (c) we will also want to solve equations
λ1 a1 + λ2 a2 + λ3 a3 = u, (? )
where u is equal to a, b, c in succession. This equation is equivalent to Ax = u, where x = [λ1 , λ2 , λ3 ]T .

We can augment A by all three vectors a, b, c and obtain all three solutions (if they exist) at the same
time by reduction of the augmented matrix to RRE form:
   
−2 3 2 | 2 1 3 1 0 0 | 0 1 1
[A|a|b|c] =  2 −2 −1 | −1 0 −1 ∼ 0 1 0 | 0 1 1 .
2 3 1 | 1 5 6 0 0 1 | 1 0 1
This is RRE form. From it we see that matrix A has rank 3.
(i) This means that rank[A|u] = 3 = rank A for any column u ∈ R3 , so that the equation Ax = u is
solvable for every u ∈ R3 ; so the set B is a spanning set for R3 .
(ii) It also means that the columns of A are linearly independent, and so the vectors comprising the set
B are linearly independent.
Hence B is a basis for R3 .
(b) (1, 1, 8), (−4, 3, 1)
(c) From the RRE form we can also read off the solutions for the three linear systems: Ax = a has a
solution x = [0, 0, 1]T , Ax = b has a solution x = [1, 1, 0]T , and Ax = c has a solution [1, 1, 1]T . Since
these equations are equivalent to (? ) for u = a, u = b, u = c, respectively, we have
     
0 1 1
[a]B = 0 , [b]B = 1 , [c]B = 1 .
1 0 1
50. (a) Independent

1 2 0 −1 1 0
(b) Dependent: +2 =
−1 0 1 1 1 2
49
MAST10018: Linear Algebra Extension Studies Answers 4: Inner product spaces
Answers 4: Inner product spaces

Inner product spaces
1. Proof required.
2. No; (i) h(0, 1), (0, 1)i = −1 < 0; (ii) h(1, 1), (1, 1)i = 0.
3. The inner product axioms are labelled (I1)–(I4).
It is convenient to write
1 −1 y1
= xT Ay.

hx, yi = x1 x2
−1 3 y2
(I1) hy, xi = y T Ax = xT AT y = xT Ay = hx, yi, as AT = A.
(I2) For u, v and y in R2 ,
hu + v, yi = (u + v)T Ay = (uT + v T )Ay

= uT Ay + v T Ay = hu, yi + hv, yi.
(I3) Similarly, hλx, yi = (λx)T Ay = λ(xT Ay) = λhx, yi.
(I4) hx, xi = x21 − 2x1 x2 + 3x22 = (x1 − x1 )2 + 2x22 > 0, for all x = (x1 , x2 ); if hx, xi = 0, then (x1 − x2 )2 = 0
and x22 = 0. So x1 = x2 = 0, and x = (0, 0).
4. (a) Yes (b) No

(c) No (d) No (e.g. h(0, 0, 1), (0, 0, 1)i = 0)
5. (a) Yes (b) No (c) Yes.
6. (a) 3 (b) −29
√ √
7. (a) kxk = 2, kyk = 1, θ = 14 π (b) kxk = 2, kyk = 3, θ = 16 π
√
8. 17
√
9. 74
√
10. 3 2
√
11. 105
12. Proof required.
13. Proof required (hint: think about what happens if Ax = 0).
14. Proof required.
Orthogonality and orthonormal bases
15. (a) Yes (b) No
1
16. (a) − √ (b) 0
2
50
17. Show hp, qi = 0.
18. (a) Yes (b) No
19. If v − w is orthogonal to v + w, then these vectors have inner product zero, i.e. 0 = hv − w, v + wi.
Expanding this product using the linearity properties we obtain
0 = hv, v + wi − hw, v + wi
= hv, vi + hv, wi − hw, vi − hw, wi.
By the symmetry property, hv, wi = hw, vi (for inner products in real spaces), so the middle terms cancel.
Also, by definition of length, hv, vi = ||v||2 and hw, wi = ||w||2 . So the last equation gives:
0 = ||v||2 − ||w||2 , or ||v||2 = ||w||2 .
Taking the positive square root of both sides gives ||v|| = ||w|| as desired.
Conversely, if ||v|| = ||w|| then we can reverse each step of this argument to show that hv − w, v + wi = 0,
i.e. v − w is orthogonal to v + w.
Geometrically: v + w and v − w represent the diagonals of a parallelogram with sides parallel to v and w,
and orthogonal vectors are perpendicular. So the result above says that the diagonals of a parallelogram
are perpendicular if and only if all the sides of the parallelogram have equal length (i.e. the parallelogram
is a rhombus).
20. (a) Proof required. (b) Proof required. (c) Proof required.
21. (a) ( 31 , − 23 , 23 ) + 2(− 32 , 13 , 23 ) + 3( 23 , 23 , 13 ) (b) 13 ( 13 , − 23 , 23 ) + 43 (− 23 , 13 , 23 ) − 31 ( 23 , 23 , 13 )
22. Proof required.
23. (a) Let S = {x1 , x2 , x3 } = {(1, 0, 1, 0), (2, 1, 1, 1), (1, −1, 1, −1)} ∈ R4 . The vectors in S are linearly
independent and span W = (S)ab . We use Gram-Schmidt orthogonalization to obtain an orthonormal
basis for W with respect to the standard (Euclidean or dot) product on R4
(u, v)ab = u · v = u1 v1 + u2 v2 + u3 v3 + u4 v4 .
√ √
Let y1 = x1 . Then ||y1 || = 1 + 1 = 2, so
y1 1
e1 = = √ (1, 0, 1, 0)
||y1 || 2
is a unit vector. Next use orthogonal projection to obtain
3 1
y2 = x2 − hx2 , e1 ie1 = (2, 1, 1, 1) − (1, 0, 1, 0) = (1, 2, −1, 2).
2 2
Then
y2 1
e2 = = √ (1, 2, −1, 2).
||y2 || 10
Continuing we calculate y3 = x3 − hx3 , e1 ie1 − hx3 , e2 ie2 , giving
2 1
y3 = (1, −1, 1, −1) − (1, 0, 1, 0) + (1, 2, −1, 2) = (2, −1, −2, −1)
5 5
and
y3 1
e3 = = √ (2, −1, −2, −1).
||y3 || 10
Hence B = {e1 , e2 , e3 } is the required orthonormal basis for W as can be verified by direct calculation.
51
1 1 1
(b) 3 (2, 2, −1, 0), 3 (0, 1, 2, −2), 3 (1, 0, 2, 2)
1 √1 (1, 1, 0, 0, −1)
(c) 4 (1, −2, 1, 3, −1), 3
so the space has dimension 2
1
24. (a) (5x − 2y + 4z, −2x + 8y + 2z, 4x + 2y + 5z)
9
1
(b) (5x + 6y + 3z, 6x + 10y − 2z, 3x + −2y + 13z)
14
(b) Proof required.
(c) Hint: Use orthogonal projections to show that any vector v in V can be written as v = w + w⊥ where
w ∈ W and w⊥ ∈ W ⊥ . Then use Vector Spaces Q44(b).
26. (a) p(x) = − 54 + 13

10 x (b) p(x) = − 74 + 1
10 x + 11 2
14 x
27. 1.55 metres
28. (a) p(x) = 303 − 86x (b) p(x) = 173 + 62.6x − 37.1x2
√ √
1 3 3 5 1
29. √ , √ x, √ x2 −
2 2 2 2 3
30. Let
1 1 1
f0 (x) = √ , fn (x) = √ sin nx, gn (x) = √ cos nx
2π π π
for n = 1, 2, 3, . . . We have to show that
hf0 , f0 i = 1, hfn , fn i = 1, hgn , gn i = 1,

hf0 , fn i = 0, hf0 , gn i = 0, hfn , fm i = 0,
hgn , gm i = 0, hfn , gk i = 0,
where m, n, k ∈ N, and m 6= n. We observe that

Z 2π Z 2π
sin px dx = 0, cos qx dx = 0, p, q ∈ Z, q 6= 0.
0 0
Calculations:
52
Z 2π 2
1
hf0 , f0 i = √ dx
0 2π
Z 2π
1
= dx = 1;
2π 0
Z 2π
1
hf0 , fn i = √ sin nx dx = 0;
π 2 0
Z 2π
1
hf0 , gn i = √ cos nx dx = 0;
π 2 0
1 2π 2
Z
hfn , fn i = sin nx dx
π 0
Z 2π
1 1
= (1 − cos 2nx) dx
π 0 2
Z 2π
1
= dx = 1;
2π 0
1 2π
Z
hgn , gn i = cos2 nx dx
π 0
1 2π 1
Z
= (cos 2nx + 1) dx
π 0 2
Z 2π
1
= dx = 1;
2π 0
1 2π
Z
hfn , fm i = sin nx sin mx dx
π 0
Z 2π
1
= (cos(n − m)x − cos(n + m)x) dx = 0;
2π 0
1 2π
Z
hgn , gm i = cos nx cos mx dx
π 0
Z 2π
1
= (cos(n + m)x + cos(n − m)x) dx = 0;
2π 0
1 2π
Z
hfn , gk i = sin nx cos kx dx
π 0
Z 2π
1
= (sin(n + k)x + sin(n − k)x) dx = 0.
2π 0
31. (a) This is a special case of (b) with k = 2.
(b) The least square approximation Af of f by trigonometric polynomials of order 6 k is given by

k
X k
X
Af = hf, fn ifn + hf, gn ign .
n=0 n=1
53
We calculate the coefficients:

1
Z 2π √
hf, f0 i = √ (1 + x) dx = 2π(1 + π);
2π 0
Z 2π
1
hf, fn i = √ (1 + x) cos nx dx
π 0
Z 2π
1
= √ x cos nx dx
π 0
2π Z 2π
1 1
= √ x sin nx − √ sin nx dx = 0;

n π 0 n π 0
Z 2π
1
hf, gn i = √ (1 + x) sin nx dx
π 0
Z 2π
1
= √ x sin nx dx
π 0
2π Z 2π √
1 1 2 π
= − √ x cos nx + √ cos nx dx = − .

n π 0 n π 0 n
So
k
X 2
Af = (1 + π) − sin nx.
n=1
n
54
MAST10018: Linear Algebra Extension Studies Answers 5: Linear transformations
Answers 5: Linear transformations

Linear transformations
1. (a) Proof required. (b) Proof required.
2. (a) Is a linear transformation. (b) Is not a linear transformation.
3. (−10, −7, 6)
4. (a) Reflection in the line y = x.
(b) Orthogonal projection onto the x-axis followed by the reflection in the line y = x.
(c) Projection onto the x-axis in the direction of the line y = −x.
(d) Shear parallel to the y-axis.
(e) Expansion by factor |b| along the x-axis and by |c| along the y-axis, with a possible change of direction
of the axis (if b < 0 or c < 0).
(f) Rotation about 0 through θ = arctan 34 .

 √ √   
− 2 − 2
0 1
5. (a)  √ 2 2
√  (b)  
2
2 − 22 −1 0
   
0 1 1 0
(c)   (d)  
1 0 0 −1
     
1
0 0 −1 0 1
6. (a) 2  (b)   (c)  
0 5 −1 0 −5 0
 
1 0 0  
  2 −1 0
7. (a) 1 1 0 (b) 
  
  1 2 0
1 1 1
 
  2 1
0 0 1 



  1 1
(c) 0 1 0 (d) 
  
−1
 
  1
1 0 0  
1 −2
8. (a) LK(x, y, z) = (x − y, 3x + 2y)
(b) T L(x, y, z) = (5x, 3x + y, x − 3y, −5y)
(c) S 2 (x, y, z) = (x, y, z)
(d) (K + S)(x, y, z) = (x + z, x + 2y, 2x + y + z)
(e) Not defined (domain T 6= range T )
55
 
5 0 0
   
−1
 
1 0 3 1 0
9. (a)   (b)  
−3
 
3 2 0 1 0
 
0 −5 0
   
1 0 0 1 0 1
   
(c) 0 (d) 1
   
1 0 2 0
   
0 0 1 2 1 1
10. Both are.
11. w1 is not; w2 is.
12. Nullity 2; basis for ker(T ) is {(−1, 1, 1, 0), (−1, 0, 0, 1)}; not injective
13. Rank 2; basis for Im (T ) is {(1, 1, 2), (2, 0, −4)}; not surjective

1 1 1 1
14. (a) (i)A = (ii) ker T = {0} (iii) , ; Im T = R2
0 3 0 3
 
 −1 
1 1 −1 1 1
(b) (i)A = (ii)  2  (iii) , ; Im T = R2
2 1 0 2 1
1
 
 
1 2
(c) (i) The standard matrix of the given linear transformation is A = 0 −1.
1 −1
 
1 0
(ii) To find a basis for the kernel of T we solve the equation Ax = 0. The RRE form for A is 0 1,
0 0
and so the equation has only the zero solution. Hence ker T = {0}.
(iii) The image of T is spanned by the columns of A. They are linearly independentas the
 reduced

 1 2 
row echelon form of A shows rank A = 2. The columns of A form a basis of Im T , ie. 0 , −1
1 −1
 
 
3 −1 −6
(d) (i) The standard matrix of T is A = −2 1 5 .
3 3 6
(ii) The kernel of T is the solution space of the equation Ax = 0.

   
3 −1 −6 1 0 −1
A = −2 1 5  ∼ 0 1 3  .
3 3 6 0 0 0
Thus {[1, −3, 1]T } is a basis for ker T .
(iii) The image of T is spanned by the columns of A. From the reduced row echelon form we see that
a possible basis can be formed from the first and second columns: {[3, −2, 3]T , [−1, 1, 3]T }.
15. (a) T −1 (x, y, z) = ( 23 x − 21 y − 12 z, x − y, − 12 x + 21 y + 12 z)
(b) Not invertible
56

cos θ sin θ
(c) Tθ−1 = T−θ =
− sin θ cos θ
(d) Tθ is its own inverse.
16. Using row operations simultaneously on the given matrix and the identity matrix we have
   
1 1 0 1 0 0 1 0 −1 1 −1 0 r1 → r1 − r2
 0 1 1 0 1 0 ∼ 0 1 1 0 1 0 
1 0 1 0 0 1 0 0 2 −1 1 1 r3 → r3 − r1 + r2
 
1 0 0 1/2 −1/2 1/2 r1 → r1 + r3 /2
∼ 0 1 0 1/2 1/2 −1/2  r2 → r2 − r3 /2
0 0 1 −1/2 1/2 1/2 r3 → r3 /2
So
 −1  
1 1 0 1 −1 1
1
0 1 1 =  1 1 −1 .
2
1 0 1 −1 1 1
   
x x−y+z
Lastly, we observe that T −1 y  = 12  x + y − z .
z −x + y + z
17. (a) 1 (b) Rotation about the z-axis through θ (c) Proof required.
 
cos θ sin θ 0
(d) Inverse of A(θ) = A(−θ) = − sin θ cos θ 0 = A(θ)T
0 0 1
 
0 0 0
18. (a) 0 0 0
0 0 1
 
4 4 2
1
(b) 4 4 2
9
2 2 1
(c) The plane x + y + z = 0 has unit length normal vector n = ( √13 , √13 , √13 ). Any vector x can be
written as a sum x = u + v, where u lies in the plane and v is parallel to n. This is because the plane
passes through the origin, and is thus a subspace. We may take a basis B = {b1 , b2 } for this subspace. If
we add n, we obtain a basis {b1 , b2 , n} for R3 . Note that n is orthogonal to both b1 and b2 . Hence
hx, ni = hu, ni + hv, ni = hv, ni,
and so
u = x − v = x − hv, nin = x − hx, nin.
In coordinates,
1
u = T (x, y, z) = (x, y, z) − (x + y + z)(1, 1, 1)
3
1
= (2x − y − z, 2y − z − x, 2z − x − y).
3
The standard matrix for this orthogonal projection T is thus
 
2 −1 −1
1
A= −1 2 −1 .
3
−1 −1 2
57
This gives
   
2 −1 −1 2 −1 −1
1
A2 = −1 2 −1 . −1 2 −1
9
−1 −1 2  −1 −1 2
6 −3 −3
1
= −3 6 −3 = A.
9
−3 −3 6
An alternative approach to show T ◦ T = T is to observe
T (T (x)) = T (u) = T (x − v) = T (x) − T (v) = T (x).
Two linear maps that agree everywhere in their domain are the same.
19.
(a) A’:(0,−.707) (b) A’:( .354,−.612) (c) A’:( −.436,−.612)
B’:(1,0) B’:( .866, .500) B’:(.433, .500)
C’:(0, .707) C’:(−.354, .612) C’:( −.789, .612)
D’:(−1,0) D’:(−.866,−.500) D’:( −.433,−.500)
E’:(0,−.707) E’:( .354,−.612) E’:(.789,−.612)
F’:(0, .707) F’:(−.354, .612) F’:(.436, .612)
y y y
C’& F’ C’& F’ F’
B’ C’
D’ B’ B’
x x D’
x
−1 1
D’ E’
A’& E’ A’
A’& E’
Matrix representations
20. (a) Given two polynomials p(x) = a + bx + cx2 , q(x) = A + Bx + Cx2 , and two scalars α, β, we have
T (αp + βq) = T (αa + βA + (αb + βB)x + (αc + βC)x2 )

= x(αa + βA + (αb + βB)x + (αc + βC)x2 ))
= αax + βAx + (αb + βB)x2 + (αc + βC)x3
= α(ax + bx2 + cx3 ) + β(Ax + Bx2 + Cx3 )
= αT (p) + βT (q)
as required for linearity.
(b) Method 1: We look at the images of the vectors in the first basis B1 = {1, x, x2 }:
T (1) = x, T (x) = x2 , T (x2 ) = x3 .
The coordinate vectors of these with respect to the second basis B2 = {1, x, x2 , x3 } give the columns
in the matrix of T with respect to the given bases. Hence, this matrix is
 
0 0 0
1 0 0
[T ] = 
0
.
1 0
0 0 1
Method 2: The matrix [T ] is defined by the property that multiplication by [T ] takes the coordinate
vector for p to the coordinate vector for T (p), i.e.
[T (p)]B2 = [T ][p]B1 for all p ∈ P2 .
58
Using the given bases we see that the polynomial p(x) = a + bx + cx2 with coordinate vector [p]
B1 =
  0
a
 b  is mapped to the polynomial T (p) = ax + bx2 + cx3 with coordinate vector [T (p)]B2 = a.
 
b
c
c
But we have    
0 0 0 0    
a 1 0 0 a a
 =   b  = [T ]  b 
 b  0 1 0
c c
c 0 0 1
so we again find that [T ] is the matrix given above.
 
1 1 1
21.  0 2 4 
0 0 4
   
0 1 0 1 0 1 0
   
   
0 0 1  0 0 0 0
22. (a)   (b)  
−1 −1 0 −2
   
0 0  0
   
0 0 0 0 0 0 0
23. (a) ker(T ) = {0}; one to one
(b) ker(T ) = {−a − 2bx + ax2 + bx3 |a, b ∈ R}; not one to one
  
 a b 
24. (a) Im (T ) =   | a, b, c ∈ R (b) Im (T ) = {a + bx2 |a, b ∈ R}
 c 0 

1 0 0 1
(b) The kernel of T has a basis , ; the image of T has a basis 1 0 , 0 1 ; rank(T ) = 2,
2 0 0 2
nullity(T ) = 2

2 0 −1 0
(c)
0 2 0 −1
 
0 0 0
1 0 0
26. 
0 1

2 0
1
0 0 3
27. x − x2 /2 + 2x3 /3
Change of basis
   
1 0 1 3 −2 3
1 
28. (a) P = −2 3 0 , P −1 = 2 2 2
10
1 2 −1 7 2 −3
   
14 −4
1  1 
(b) (i) 6 (ii) 9
5 5
1 −6
29.
59
 
1 2 1 1 1 −1
; 1 (1, −5)
4 2 −1 4 (a) −1
(b) −1 2 ; (1, 0, −2)
−1 0 1
 
1 0 4 6 −4
30. (a) P −1 AP =
2 −1 (b) P −1 AP = −1 −5 5 
−2 −3 1
 
4 5 −4
31. P −1 AP = −4 −6 5
−1 −3 4

1 −2 1 −3 −56
32. ,
0 −1 11 −2 3
 
−28 −19 −43
33.  5 4 7 
18 11 28
Cayley-Hamilton Theorem
34. (a) Proof required. (b) Proof required.

 
3 −3 1
35. 1 0 0
0 1 0
36. (a) t2 + t − 11 (b) (t − 1)(t2 − 2t − 5)
60
MAST10018: Linear Algebra Extension Studies Answers 6: Eigenvalues and eigenvectors
Answers 6: Eigenvalues and eigenvectors
We reserve u1 , u2 , . . . for eigenvectors, λ1 , λ2 , . . . for eigenvalues, x1 , x2 , . . . for components of vector x; and write
W (λ) for the eigenspace corresponding to λ.
Eigenvectors, eigenvalues and diagonalisation,

0
1. (a) λ = 2, u =
1

1 2
(b) λ1 = 1, u1 = ; λ2 = 2, u2 =
3 5
(c) The characteristic polynomial is

1−λ −1 = (1 − λ)(3 − λ) + 1 = λ2 − 4λ + 4.

1 3−λ
The eigenvalues are therefore both 2. To find a basis for the corresponding eigenspace, we solve the
equation
x −1 −1 x 0
(A − 2I) = = .
y 1 1 y 0

1
This gives y = −x, and so the eigenspace is 1-dimensional with basis .
−1

i −i
(d) complex eigenvalues and eigenvectors: λ1 = i, u1 = , λ2 = −i, u2 = .
1 1
* + * + * + * +
1 0 2 1
2. (a)   and   (b)   and  
2 1 −1 −3
3. Rotation through θ.
For θ = kπ, k odd, Aθ = −I and for θ = kπ, k even, Aθ = I. In both cases all lines are invariant.
4. (a) 1, 3, 4 (b) 1, 3, 6, 10
5. Proof required
     
1 0 −3
6. (a) λ1 = 2, u1 = 0 ; λ2 = 2, u2 = 2 ; λ3 = 3, u3 =  3 
0 1 1
   
1 0
(b) λ1 = 2, u1 = 0 ; λ2 = −3, u2 = 0
0 1
    
1 −2 0
(c) λ1 = −1, u1 = −2 ; λ2 = 1, u2 =  0  ; λ3 = −2, u3 =  3 
1 1 −2
61
(d) The characteristic polynomial of A is p(λ) = det (A − λI) which may be factorized directly as follows

2−λ 2 2

det(A − λI) = −1 −1 − λ −2
1 2 3−λ

2−λ 2 2

= 0 1 − λ 1 − λ r2 → r2 + r3
1 2 3−λ

2−λ 2 2

= (1 − λ) 0 1 1 (taking out common
1 2 3−λ factor of r2 )

2−λ 0 0

= (1 − λ) 0 1 1 r1 → r1 − 2r2
1 2 3−λ

1 1
= (1 − λ)(2 − λ) (cofactor expansion of
2 3−λ
determinant by the first row)
= (1 − λ)2 (2 − λ).
So the eigenvalues are λ1 = λ2 = 1, λ3 = 2.
We first solve the eigenvector equation (A − I)x = 0 for W (1), the eigenspace of λ1 , λ2 . Reducing
A − I to RRE form    
1 2 2 1 2 2
A − I = −1 −2 −2 ∼ 0 0 0
1 2 2 0 0 0
we find the eigenvector equation is equivalent to
x1 + 2x2 + 2x3 = 0.
Choosing the non-leading variables, x2 , x3 , as arbitrary parameters, x1 = −2x2 − 2x3 . So if u = [x1 , x2 , x3 ]T

is a vector in W (1) then
     
−2x2 − 2x3 −2 −2
u= x2  = x2  1  + x3  0 
x3 0 1
Thus we may choose {u1 = [−2, 1, 0]T , u2 = [−2, 0, 1]T } as a basis for W (1).
Next we solve the eigenvector equation (A − 2I)x = 0. Reducing the matrix A − 2I to RRE form
   
0 2 2 1 0 −1
A − 2I = −1 −3 −2 ∼ 0 1 1 
1 2 1 0 0 0
we find the eigenvector equation is equivalent to
x1 − x3 = 0,
x2 + x3 = 0.
Choosing the non-leading variable x3 as an arbitrary parameter, we have x1 = x3 and x2 = −x3 . So
if u = [x1 , x2 , x3 ]T is a vector in W (2) then
   
x3 1
u = −x3  = x3 −1
x3 1
So a basis for W (2) is {[1, −1, 1]T }.
Thus A has eigenvalues λ1 = λ2 = 1 and λ3 = 2 with corresponding eigenvectors u1 = [−2, 1, 0]T ,

u2 = [−2, 0, 1]T , u3 = [1, −1, 1]T .
62
7. We write W (λ) for the eigenspace corresponding to λ. The spanning sets for W (λ) listed below are bases.
*−1 −1+ *1+
(a) λ1 = λ2 = 2, W (2) =  1  ,  0  ; λ3 = 6, W (6) = 2
0 1 1
*1 0+
(b) λ1 = λ2 = λ3 = 1, W (1) = 0 , 0
0 1
8. Question 1:
(a) not diagonalizable

1 2 −5 2 1 0
(b) diagonalizable; P = , P −1 = , D=
3 5 3 −1 0 2
(c) The matrix is not diagonalizable, since we cannot find a basis for R2 consisting of eigenvectors of A.
(d) diagonalizable
(using complex
D,
P );
i −i −1 1 −i 1 i 0
P = , P =2 , D=
1 1 i 1 0 −i
Question 6:
(a) diagonalizable;
     
1 0 −3 1 3 −6 2 0 0
P = 0 2 3  , P −1 = 0 −1 3  , D = 0 2 0
0 1 1 0 1 −2 0 0 3
(b) not diagonalizable
(c) diagonalizable;
     
1 −2 0 3 4 6 −1 0 0
P = −2 0 3  , P −1 = 1 2 3 , D =  0 1 0
1 1 −2 2 3 4 0 0 −2
(d) The matrix has three linearly independent eigenvectors, and therefore is diagonalizable. Form a
matrix P whose columns are the eigenvectors u1 , u2 , u3 , and a matrix Λ = diag(1, 1, 2). Then
     
−2 −2 1 1 0 0 1 3 2 2 2 2
P ΛP −1 =  1 0 −1 0 1 0 −1 −2 −1 = −1 −1 −2 .
0 1 1 0 0 2 1 2 2 1 2 3
9. This problem reduces to the observation that the product of two diagonal matrices is a diagonal matrix,
with entries the products of the corresponding entries of the matrices being multiplied. In other words, if
D = diag(λ1 , . . . , λn ) and E = diag(µ1 , . . . , µn ), then DE = diag(λ1 µ1 , . . . , λn µn ) = ED.
The argument is as follows:

Eij = 0 if i 6= j, and Eii = µi , and similarly for D. Hence to calculate the (i, j) entry of DE, we see
n
X
(DE)ij = Dik Ekj = Dii Eij .
k=1
The last term vanishes unless i = j, and so the only non-zero terms possible are those on the diagonal of
form Dii Eii = λi µi .
We can now perform the induction.

STEP 1. For n = 1, this is the definition of D.
63
STEP 2. Suppose the formula is true for n = k. We calculate

Dk+1 = DDk = diag(λ1 , . . . , λn ) · diag(λk1 , . . . , λkn ).
Since this is the product of two diagonal matrices, we obtain
Dk+1 = diag(λ1 λk1 , . . . , λn λkn ) = diag(λk+1
1 , . . . , λk+1
n ).
The formula is thus true for n = k + 1, and so by induction, it is true for all positive integers n.
10. STEP 1. For n = 1 the equation holds as it becomes the hypothesis A = P DP −1 .
STEP 2. Suppose the equation is true for n = k. Then

Ak+1 = AAk = (P DP −1 )(P Dk P −1 )
= (P D)(P P −1 )(Dk P ) = P DIDk P −1 = P Dk+1 P −1 ,
i.e. the equation is true for n = k + 1, if it is true for n = k.
So by the principle of mathematical induction the equation is true for all integers n > 1, i.e. the equation
is true for all n ∈ N.
11. (a) The characteristic equation of A is p(λ) = λ2 − λ − 2 = (λ − 2)(λ + 1) = 0. So the eigenvalues are
λ1 = 2, λ2 = −1. The corresponding eigenvector equations are (A − 2I)x = 0 and (A + I)x = 0. After
reduction they become x1 − 2x2 = 0 (for λ1 = 2), and 2x1 − x2 = 0 (for λ2 = −1). Corresponding
eigenvectors are u1 = [2, 1]T , u2 = [1, 2]T . So A is diagonalizable with

2 1 2 0 1 2 −1
A = P DP −1 = .
1 2 0 −1 3 −1 2
Therefore A5 = P D5 P −1 by the preceding question, and

5 1 2 1 32 0 2 −1 1 129 −66 43 −22
A = = = .
3 1 2 0 −1 −1 2 3 66 −36 22 −12
 
1509 198 −1464
(b) A5 = P D5 P −1 = 1477 230 −1464
1477 231 −1465
 
8 31 357 155
0 258 254 −6 
(c) A5 = P D5 P −1 = 18 
0 62 −318

−186
0 68 188 52
" #
1
2
12. Proof is required. (Hint: First diagonalize the matrix; eigenvalues are 1, 1/5). lim xn = 1
.
n→∞
2
13. The Fibonacci numbers 0, 1, 1, 2, 3, 5, 8, 13, . . . are defined recursively by the equation
Fk+2 = Fk+1 + Fk

F
and the initial conditions F0 = 0, F1 = 1. Putting uk = k+1 , equation (∗) gives:
Fk

F 1 1 Fk+1 1
uk+1 = k+2 = = Auk and u0 = .
Fk+1 1 0 Fk 0
Hence, by induction on k, we find u1 = Au0 , u2 = Au1 = A(Au0 ) = A2 u0 , . . . uk = Ak u0 . Now A has
√ √
λ1 λ
eigenvalues λ1 = 1+2 5 , λ2 = 1−2 5 with corresponding eigenvectors , 2 . Hence A = P DP −1
1 1
where
λ 0 λ 1 1 1 −λ2
D= 1 , P = 1 , P −1 = √ .
0 λ2 λ2 1 5 −1 λ1
64
Then
Fk+1
= uk = Ak u0 = P Dk P −1 u0
Fk
and after some algebra this gives
√ !k √ !k
1 1 1+ 5 1− 5
Fk = √ λk1 − λk2 = √

− .
5 5 2 2
Now
Fk+1 λk+1 − λk+1 λ1 − λ2 ( λλ21 )k
lim = lim 1 k 2
= ,
k→∞ Fk k→∞ λ1 − λk2 1 − ( λλ21 )k

λ2
after dividing top and bottom by λk1 . Since λ1 ≈ 1.618 and λ2 ≈ −0.618, we have < 1, so
λ1
k
λ2
lim = 0. Hence
k→∞ λ1
√
Fk+1 1+ 5
lim = λ1 = = “golden ratio” ≈ 1.618 . . .
n→∞ Fk 2
14. Write xi = Lemon’s share and yi = LIME’s share after i years.
(a) xi+1 = 0.85xi + 0.25yi , yi+1 = 0.15xi + 0.75yi with x0 = 0.60 and y0 = 0.40.

xi x .85 .25
(b) = Ai 0 where A = .
yi y0 .15 .75

0.6231 0.6248
(c) , (d) Lemon 62.5%, LIME 37.5%.
0.3769 0.3752
15. Proofs required.
16. (a) Orthogonal (b) Not orthogonal
17. Proof required.

   
2 −2 1 9 0 0
18. (a) P = 13 2 1 −2 , D = 0 6 0
1 2 2 0 0 3
   
0 4 3 −3 0 0
(b) P = 15 5 0 0 , D= 0 25 0 
0 −3 4 0 0 −50
 √1 √1 0
  
2 2 2 0 0
(c) P =  √1 − √12 0 , D = 0 0 0
2
0 0 1 0 0 0
(d) The characteristic polynomial of A is p(λ) = det(A − λI), which may be factorized directly as follows

4−λ 2 2

det(A − λI) = 2 4−λ 2
2 2 4−λ

2−λ 0 −2 + λ
2 − λ −2 + λ r1 → r1 − r3

= 0
2 2 4 − λ r2 → r2 − r3
65

1 0 −1
= (2 − λ)2

0 1 −1
(taking out common

2 2 4−λ
factors of r1 and r2 )

1 0 0
= (2 − λ)2

0 1 −1 c3 → c3 + c1
2 2 6−λ

2
1 −1 (cofactor expansion
= (2 − λ)
2 6−λ by the first row)
= (2 − λ)2 (8 − λ).
We first solve (A − 2I)x = 0, the eigenvector equation for the eigenvalue λ1 = λ2 = 2. It is equivalent
to x1 + x2 + x3 = 0. Choosing x2 and x3 as arbitrary parameters, and if u = [x1 , x2 , x3 ]T is a vector
in W (2) then      
−x2 − x3 −1 −1
u= x2  = x2  1  + x3  0  .
x3 0 1
Thus we may choose as a basis for W (2) the set {u1 = [−1, 1, 0]T , u2 = [−1, 0, 1]T }.
Reducing A − 8I to RRE form we find the eigenvector equation (A − 8I)x = 0 is equivalent to

x1 − x3 = 0, x2 − x3 = 0. Choosing x3 to be an arbitrary parameter, and if u = [x1 , x2 , x3 ]T a
vector in W (8) then    
x3 1
u = x3  = x3 1 .
x3 1
So a basis for W (8) is {[1, 1, 1]T }.
As we want an orthogonal matrix, P , the vectors u1 , u2 must be orthogonalized first. This can be
done in two ways:
(i) u2 is replaced by v2 = u1 × u3 ;
(ii) u2 is replaced by v2 = u1 + tu2 , where t is chosen to make v2 orthogonal to u1 (a special case

of Gram-Schmidt orthogonalization).
Method (ii) is preferred as it generalizes easily to higher dimension. Proceeding with method (ii) we
have
0 = u1 · v2 = |u1 |2 + tu1 · u2 = 2 + t.
So v2 = u1 − 2u2 = [1, 1, −2]T . We choose the columns of the matrix P to be the normalized vectors
u1 , v2 , u3 . So
 1
√1 √1
   √1 √1

− √2 − 0
 
6 3 2 0 0 4 2 2
 √1 2 2
P DP T =  √12 √1 √1  0 2 0  6 √1 − √26  = 2 4 2 ,
 
6 3  6 
0 − √26 √13 0 0 8 √1
3
√1
3
√1
3
2 2 4
which is A.
19. Proof required.
20. Proof required.
66

MAST10018 Exercises

Uploaded by

Document Information

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

MAST10018 Exercises

Uploaded by

Copyright:

Available Formats

MAST10018

© University of Melbourne 2016

Exercises 2: Vectors and geometry 7

Exercises 3: Vector spaces 11

Exercises 4: Inner product spaces 18

Exercises 5: Linear transformations 22

Exercises 6: Eigenvalues and eigenvectors 28

Answers 1: Linear equations and matrices 31

Answers 2: Vectors and geometry 39

Answers 3: Vector spaces 44

Answers 4: Inner product spaces 50

Answers 5: Linear transformations 55

Answers 6: Eigenvalues and eigenvectors 61

Content last updated: January 4, 2018

Exercises 1: Linear equations and matrices

1. Evaluate the following matrix products:

5. Give 3 × 3 examples of the following:

(a) a diagonal matrix; that is, a matrix A with Aij = 0 if i 6= j.

(c) a symmetric matrix; that is, a matrix A with Aij = Aji .

6. Use trial and error to find 2 × 2 examples of the following:

(a) a non-zero matrix A with A2 = 0;

(b) a matrix B with real entries and with B 2 = −I;

(c) matrices C and D with no zero entries but with CD = 0;

(d) matrices E and F with EF = −F E but EF 6= 0.

7. Suppose that a 2 × 2 matrix A satisfies AB = BA for every 2 × 2 matrix B. Set

Then AB = BA when B is either of

9. Which of the following systems of equations are linear?

(a) x1 − 3x2 = x3 − 4 (b) 3x − xy =1

The row reduction algorithm

Homogeneous linear equations

13. Using row reduction, solve the following homogeneous systems:

(a) (2 + i)x − iy + (3 − i)z + u = 0 (b) 7x + 7y − 16z + 6w = 0

Inhomogeneous linear equations

(a) kx + y + z = 1 (b) 2x + (k − 4)y + (3 − k)z = 1

For the cases (ii) and (iii) find the solutions.

17. Determine the conditions on a, b, c so that the system has a solution:

Find the solutions when they exist.

19. Repeat the preceding problem for the complex matrices

20. Find the inverse of the matrix  

21. Evaluate the following determinants by reduction to triangular form:

23. Using a suitable determinant find

24. Establish the following factorizations:

25. Evaluate the following determinants:

26. The famous Vandermonde determinant is defined by

27. Idempotent Matrices A matrix P is called idempotent if P 2 = P . If P is idempotent and P 6= I show

Exercises 2: Vectors and geometry

1. For the given vectors a, b find a + b, 5a − 4b, a · b, kak, and ka − bk.

(a) a = (−2, 6, 1), b = (3, −3, −1)

(b) a = 3e1 − 4e2 + 21e3 , b = e1 + 2e2 − 5e3

2. Find the following scalar products:

3. Find the angle between the following pairs of vectors:

5. Find the following vector products:

(a) (e1 − e2 − 2e3 ) × (2e1 + e2 − 3e3 ) (b) (2, 1, 1) × (1, −3, 7)

6. Find unit vectors orthogonal to the following pairs of vectors:

Some vector identities

9. If a · b = 0 and x + (x · b)a = b, find the vector x in terms of a and b.

10. Show that a · (b × c) = c · (a × b) = −b · (a × c).

11. Prove that a × (b × c) = (a · c)b − (a · b)c.

12. Use Qs 10 and 11 to show that (a × b) · (c × d) = (a · c) (b · d) − (a · d) (b · c).

Deduce Lagrange’s identity: ka × bk2 = kak2 kbk2 − (a · b)2 .

13. Prove the following vector identities:

Hint. Use Qs 11 and 12.

14. Prove the following properties of the vector product:

(a) a × b = −b × a (b) (αa + βb) × c = αa × c + βb × c