Copies available at http://www.damtp.cam.ac.

uk/user/examples/

A1a Vectors and Matrices: Example Sheet 1 Michaelmas 2014

A ∗ denotes a question, or part of a question, that should not be done at the expense of questions on later sheets.
Starred questions are not necessarily harder than unstarred questions.
Corrections and suggestions should be emailed to [email protected].

1. Let S be the interior of the circle |z − 1 − i| = 1. Show, by using suitable inequalities for |z1 ± z2 |,
that if z ∈ S then √ √
5 − 1 < |z − 3| < 5 + 1.
Obtain the same result geometrically by considering the line containing the centre of the circle and
the point 3.
2. Given |z| = 1 and arg z = θ, find both algebraically and geometrically the modulus-argument forms
of
Copyright © 2014 University of Cambridge. Not to be quoted or reproduced without permission.

(i) 1 + z, (ii) 1 − z .
Show that the locus of w as z varies with |z| = 1, where w is given by
µ ¶
1−z
w2 = ,
1+z
is a pair of straight lines.
3. Use complex numbers to show that the medians of a triangle are concurrent.
Hint: represent the vertices of the triangle by complex numbers z1 , z2 and z3 (or 0, z1 and z2 if you
prefer), then write down equations for two of the medians and find their intersection.
∗4. Express
z5 − 1
I=
z−1
as a polynomial in z. By considering the complex fifth root of unity ω, obtain the four factors of I
linear in z. Hence write I as the product of two real quadratic factors. By considering the term in z 2
in the identity so obtained for I, show that
π π
4 cos sin =1.
5 10

5. Show the equation sin z = 2 has infinitely many solutions.

6. (a) Let z, a, b ∈ C (a 6= b) correspond to the points P, A, B of the Argand plane. Let Cλ be the locus
of P defined by
P A/P B = λ ,
where λ is a fixed real positive constant. Show that Cλ is a circle, unless λ = 1, and find its
centre and radius. What if λ = 1?
∗(b) For the case a = −b = p, p ∈ R, and for each fixed µ ∈ R, show that the curve
n p o
Sµ = z ∈ C : |z − iµ| = p2 + µ2

is a circle passing through A and B with its centre on the perpendicular bisector of AB.
Show that the circles Cλ and Sµ are orthogonal for all λ, µ.
7. Show by vector methods that the altitudes of a triangle are concurrent.
Hint: let the altitudes AD, BE of △ABC meet at H, and show that CH is perpendicular to AB.
8. Given that vectors x and y satisfy
x + y(x · y) = a ,
for fixed vector a, show that
|a|2 − |x|2
(x · y)2 = .
2 + |y|2

Mathematical Tripos IA: Vectors and Matrices 1 [email protected], Michaelmas 2014

 Copies available at http://www.damtp.cam.ac.uk/user/examples/

Deduce using the Schwarz inequality (or otherwise) that

|x|(1 + |y|2 ) > |a| > |x| .
Explain the circumstances under which either of the inequalities can be replaced by equalities, and
describe the relation between x, y and a in these circumstances.
−−→ −−→ −→
9. (a) In △ABC, let AB, BC and CA be denoted by u, v and w. Show that
u × v = v × w = w × u, (∗)
and hence obtain the sine rule for △ABC.
(b) Given any three vectors p, q, r such that
p × q = q × r = r × p, (†)
with |p × q| =
6 0, show that
Copyright © 2014 University of Cambridge. Not to be quoted or reproduced without permission.

p + q + r = 0.
10. (a) Using the identity a × (b × c) = (a · c)b − (a · b)c, deduce that
(i) (a × b) · (c × d) = (a · c)(b · d) − (a · d)(b · c),
(ii) a × (b × c) + b × (c × a) + c × (a × b) = 0.
Relate the case c = a, d = b of (i) to a well-known trigonometric identity.
Evaluate (a × b) × (c × d) in two distinct ways and use the result to display explicitly a linear
dependence relation amongst the four vectors a, b, c and d.
(b) Given [a, b, c] ≡ a · (b × c), show that
[a × b, b × c, c × a] = [a, b, c]2 .

∗11. For φ, θ ∈ R, let the vectors er , eθ and eφ in R3 be defined in terms of the Cartesian basis (i, j, k) by
er = cos φ sin θ i + sin φ sin θ j + cos θ k ,
eθ = cos φ cos θ i + sin φ cos θ j − sin θ k ,
eφ = − sin φ i + cos φ j .
Show that (er , eθ , eφ ) constitute an orthonormal right-handed basis. Discuss the significance of this
[local] basis.
12. The set X contains the six real vectors
(1, 1, 0, 0), (1, 0, 1, 0), (1, 0, 0, 1), (0, 1, 1, 0), (0, 1, 0, 1), (0, 0, 1, 1).
Find two different subsets Y of X whose members are linearly independent, each of which yields a
linearly dependent subset of X whenever any element x ∈ X with x ∈
/ Y is adjoined to Y .
13. Let V be the set of all vectors x = (x1 , . . . , xn ) in Rn (n > 4) such that their components satisfy
xi + xi+1 + xi+2 + xi+3 = 0 for i = 1, 2, . . . , n − 3 .
Find a basis for V .
14. Let x and y be non-zero vectors in Rn with scalar product denoted by x · y. Prove that
(x · y)2 6 (x · x)(y · y) ,
and prove also that (x · y)2 = (x · x)(y · y) if and only if x = λy for some scalar λ.
(a) By considering suitable vectors in R3 , or otherwise, prove that the inequality
x2 + y 2 + z 2 > yz + zx + xy
holds for any real numbers x, y and z.
(b) By considering suitable vectors in R4 , or otherwise, show that only one choice of real numbers x,
y and z satisfies
3(x2 + y 2 + z 2 + 4) − 2(yz + zx + xy) − 4(x + y + z) = 0 ,
and find these numbers.

Mathematical Tripos IA: Vectors and Matrices 2 [email protected], Michaelmas 2014

 Copies available at http://www.damtp.cam.ac.uk/user/examples/

A1b Vectors and Matrices: Example Sheet 2 Michaelmas 2014

Corrections and suggestions should be emailed to [email protected].

1. In the following, the indices i, j, k, l take the values 1, 2, 3, and the summation convention applies. In
particular, ni ni = 1; i.e., ni are the components of a unit vector n.

(a) Simplify the following expressions:

δij aj , δij δjk , δij δji , δij ni nj , εijk δjk , εijk εijl , εijk εikj , εijk (a × b)k .

(b) Given that Aij = εijk ak (for all i, j), show that 2ak = εkij Aij (for all k).
(c) Show that εijk sij = 0 (for all k) if and only if sij = sji (for all i, j).
(d) Given that Nij = δij − εijk nk + ni nj and Mij = δij + εijk nk , show that Nij Mjk = 2δik .
Copyright © 2014 University of Cambridge. Not to be quoted or reproduced without permission.

2. Let a, b, c and d be fixed vectors in R3 . In each of cases (i) and (ii) find all vectors r such that
(i) r + r × d = c , (ii) r + (r · a) b = c .
In (ii) consider separately the a · b 6= −1 and a · b = −1 subcases.
Hint: given r0 solving (ii) for a·b = −1, show that r0 +λb is another solution for an arbitrary scalar λ.
3. In R3 show that the straight line through the points a and b has equation
r = (1 − λ)a + λb,
and that the plane through the points a, b and c has the equation
r = (1 − µ − ν)a + µb + νc,
where λ, µ and ν are scalars. Obtain forms of these equations that do not involve λ, µ, ν.
4. (a) Let λ be a scalar, and let m, u and a be fixed vectors in R3 such that m · u = 0 and a · u 6= 0.
Show that the straight line r × u = m meets the plane r · a = λ in the point
a × m + λu
r = .
a·u
Explain in detail the geometrical meaning of the condition a · u 6= 0.
(b) In R3 show that if r lies in the planes r · a = λ and r · b = µ, for fixed non-zero vectors a and b,
and scalars λ and µ, show that
r × (a × b) = µa − λb . (∗)
Conversely, given a × b 6= 0, show that (∗) implies both r · a = λ and r · b = µ. Hence deduce that
the intersection of two non-parallel planes is a line. Comment on the case in which a × b = 0.
5. Let n be a unit vector in R3 . Identify the image and kernel (null space) of each of the following linear
maps R3 → R3 :
(a) T : x 7→ x′ = x − (x · n) n , (b) Q : x 7→ x′ = n × x .
Show that T 2 = T and interpret the map T geometrically. Interpret the maps Q2 and Q3 + Q, and
show that Q4 = T .
6. Give a geometrical description of the images and kernels of each of the linear maps of R3
(a) (x, y, z) 7→ (x + 2y + z, x + 2y + z, 2x + 4y + 2z),
(b) (x, y, z) 7→ (x + 2y + 3z, x − y + z, x + 5y + 5z).
7. A linear map A : R4 → R4 is defined by x 7→ Ax where
 
a a b a
a a b 0 
A =  a b a b  .


a b a 0
Find the kernel and image of A for all real values of a and b.

Mathematical Tripos IA: Vectors and Matrices 1 [email protected], Michaelmas 2014

 Copies available at http://www.damtp.cam.ac.uk/user/examples/

8. A linear map S : R2 → R2 is defined by

x 7→ x′ = x + λ(b · x) a ,
where λ is a scalar, and a and b are fixed, orthogonal unit vectors. By considering its effect on the
vectors a and b, show that S describes a shear in the direction of a. Let S(λ, a, b) be the matrix
with entries Sij such that x′i = Sij xj . Obtain an expression for Sij in terms of the components of a
and b and hence find the matrix S(λ, a, b). Evaluate its determinant‡ , and hence deduce that S is an
area-preserving map.
9. The linear map R3 → R3 defined by
x 7→ x′ = cos θ x + (x · n) (1 − cos θ) n − sin θ (x × n) (†)
describes a rotation by angle θ in a positive sense about the unit vector n. Verify this by considering
the case of n = (0, 0, 1).
Copyright © 2014 University of Cambridge. Not to be quoted or reproduced without permission.

Show that (†) can be written in matrix form as

x 7→ x′ = R(n, θ) x
where R(n, θ) is a matrix with entries Rij which you should find explicitly in terms of δij , εijk , etc.
Hence show that
Rii = 2 cos θ + 1 , εijk Rjk = −2ni sin θ.
Given that R(n, θ) is the matrix  
2 −1 2
1 
3
2 2 −1 ,
−1 2 2
determine θ and n.
10. Give examples of 2 × 2 real matrices representing the following transformations in R2 : (a) reflection,
(b) dilatation (enlargement), (c) shear, and (d) rotation. Which of these types of transformation are
always represented by a 2 × 2 matrix with determinant +1?
If maps A and B are both shears, will AB be the same as BA in general? Justify your answer.
11. Suppose that A and B are both Hermitian matrices. Show that AB + BA is Hermitian. Also show that
AB is Hermitian if and only if A and B commute.
∗12. Let R(n, θ) be the matrix defined by the linear map (†) of question 9, and let i, j, k be the standard
mutually orthogonal unit vectors in R3 .
(a) Show that the matrix R(i, π2 )R(j, π2 ) is orthogonal, has determinant one, and is not equal to the
matrix R(j, π2 )R(i, π2 ).
(b) Reflection in a plane through the origin in R3 , with unit normal n, is a linear map such that
x 7→ x′ = x − 2(x · n) n.
In matrix notation x′ = H(n) x for matrix H(n). Show by geometrical and algebraic means that
the map x 7→ x′ = −H(n)x, describes a rotation of angle π about n.
(c) A vector x has components (x, y, z) in a (Cartesian) coordinate system S. It has components
(x′ , y ′ , z ′ ) in a coordinate system S ′ obtained from S by anti-clockwise rotation through angle α
about axis k. Show, geometrically, that the components in coordinate system S ′ are related to
those in S by  ′  
x x
y ′  = R(k, −α) y  .
z′ z
(d) Given that ¡1 ¢ ¡1 ¢
n± = cos 2 θ i ± sin 2 θ j ,
prove that
H (i) H (n− ) = H (n+ ) H (i) = R (k, θ) ,
and give diagrams to exhibit the geometrical meaning of this result.
‡ You may need to return to this question if determinants have not been covered yet.

Mathematical Tripos IA: Vectors and Matrices 2 [email protected], Michaelmas 2014

 Copies available at http://www.damtp.cam.ac.uk/user/examples/

A1c Vectors and Matrices: Example Sheet 3 Michaelmas 2014

A ∗ denotes a question, or part of a question, that should not be done at the expense of questions on later sheets.
Starred questions are not necessarily harder than unstarred questions.
Corrections and suggestions should be emailed to [email protected].

1. Given that A is the real matrix  

a a2 bc
 b b2 ca ,
c c2 ab
show with the aid of row operations that

det A = (a − b)(b − c)(c − a)(ab + bc + ca).

Copyright © 2014 University of Cambridge. Not to be quoted or reproduced without permission.

[Recall that the value of the determinant is unchanged if a linear combination of any two rows is
added to the third row.]

2. Show that ¯ ¯
¯x y z ¯
¯ ¯
¯ z x y ¯ = x3 + y 3 + z 3 − 3xyz ≡ ∆.
¯ ¯
¯ y z x¯

Show, by row operations, that

x + y + z, x + ωy + ω 2 z, x + ω 2 y + ωz

are factors of ∆, where ω is a complex cube root of unity. Show, by considering the coefficients of
x3 , that ∆ is equal to the product of the three indicated factors.

3. If A is a (2n + 1) × (2n + 1) antisymmetric matrix (n ∈ N), calculate det A.

4. Let D be the n × n matrix which has the entry p, p 6= 1, at each place on the main diagonal and
unity in every other position. Show that det D = (p + n − 1)(p − 1)n−1 .

5. Identify the cofactors ∆ij of aij in the matrix

 
1 1 1
A = {aij } = 1 2 3 .
3 −2 2

Verify the identity aij ∆ik = δjk det A, and hence construct the matrix A−1 .
Use your result to solve the equations

x+y+z = 1,
x + 2y + 3z = −5 ,
3x − 2y + 2z = 4.

Verify that your answers for (x, y, z) do indeed satisfy the equations.

6. For each real value of t, determine whether or not there exist solutions to the simultaneous equations

x+y+z = t
tx + 2z = 3
3x + ty + 5z = 7,

exhibiting the most general form of such solutions when they exist.

Mathematical Tripos IA: Vectors and Matrices 1 [email protected], Michaelmas 2014

 Copies available at http://www.damtp.cam.ac.uk/user/examples/

∗7. Let A be a real 3 × 3 matrix, and let d be a 3 component column vector. Explain briefly how the
general solution of the matrix equation Ax = d, where x is a 3 component column vector, depends
on the kernel and image of the linear map x 7→ Ax.
Consider the case     
1 1 1 x 1
A = 1 a b , x = y  , d = −1 .
1 a2 b2 z 1
Find the kernel and image of the corresponding map, noting the different possibilities according to
different values of a and b.
For which values of a and b do the equations Ax = d have (i) a unique solution, (ii) more than
one solution, (iii) no solution? For each pair (a, b) satisfying (ii), give the solutions as the sum of
a fixed solution and the general solution of the corresponding homogeneous equations.
8. Find the eigenvalues and eigenvectors of the matrix
Copyright © 2014 University of Cambridge. Not to be quoted or reproduced without permission.

 
1 α 0
A = β 1 0 ,
0 0 1
where neither of the complex constants α and β vanishes. Find the conditions for which (a) the
eigenvalues are real, and (b) the eigenvectors are orthogonal. Hence show that both conditions are
jointly satisfied if and only if A is Hermitian.
Recall both that the scalar product for two vectors u, v ∈ C3 is defined as
u · v = u∗1 v1 + u∗2 v2 + u∗3 v3 ,
where ∗
denotes a complex conjugate, and that u and v are said to be orthogonal if u · v = 0.
9. (a) Find a 3 × 3 real matrix with eigenvalues 1, i, −i. Hint: think geometrically.
(b) Construct a 3 × 3 non-zero real matrix which has all three eigenvalues zero.
10. (a) Let A be a square matrix such that Am = 0 for some integer m. Show that every eigenvalue of
A is zero.
(b) Let A be a real 2 × 2 matrix which has non-zero non-real eigenvalues. Show that the non-
diagonal elements of A are non-zero, but that the diagonal elements may be zero.
11. Let Q be a (2n + 1) × (2n + 1) orthogonal matrix (n ∈ N) with det Q = 1. Show that Q has a unit
eigenvalue. Give a geometric interpretation of your result for 3 × 3 matrices.
∗12. Suppose that A is an n × n square matrix and that A−1 exists. Show that if A has characteristic
equation a0 + a1 t + · · · + an tn = 0, then A−1 has characteristic equation
(−1)n det(A−1 )(an + an−1 t + · · · + a0 tn ) = 0.

Hints. Take n = 3 in this question if you wish, but treat the general case if you can. It should be
clear that λ is an eigenvalue of A if and only if 1/λ is an eigenvalue of A−1 , but this result says
more than this (about multiplicities of eigenvalues). You should use properties of the determinant
to solve this problem, for example, det(A) det(B) = det(AB). You should also state explicitly why
we do not need to worry about zero eigenvalues.
13. For each of the three matrices below,
(a) compute their eigenvalues (as often happens in exercises and seldom in real life each eigenvalue
is a small integer);
(b) for each real eigenvalue λ compute the dimension of the eigenspace {x ∈ R3 : Ax = λx};
(c) determine whether or not the matrix is diagonalizable as a map of R3 into itself.
     
5 −3 2 1 −3 4 7 −12 6
6 −4 4 , 4 −7 8 , 10 −19 10 .
4 −4 5 6 −7 7 12 −24 13

Mathematical Tripos IA: Vectors and Matrices 2 [email protected], Michaelmas 2014

 Copies available at http://www.damtp.cam.ac.uk/user/examples/

A1d Vectors and Matrices: Example Sheet 4 Michaelmas 2014

A ∗ denotes a question, or part of a question, that should not be done at the expense of questions later on the sheet.
Starred questions are not necessarily harder than unstarred questions.
Corrections and suggestions should be emailed to [email protected].

1. A matrix A is said to be upper triangular if Aij = 0 if i > j, i.e. if

 
A11 A12 . . . A1 n
.
A22 . . A2 n 
 
 0
A=  . ..  .

 .. .. ..
. . . 
0 0 . . . Ann
Copyright © 2014 University of Cambridge. Not to be quoted or reproduced without permission.

Show that the eigenvalues are λi = Aii (i = 1, . . . , n, and obviously no sum).

2. Let {e1 , . . . , em } and {f1 , . . . , fn } be bases for Rm and Rn respectively, and let A be a linear mapping
from Rm to Rn . Explain how to represent A by a matrix relative to the given bases.

(a) Taking m = 2, n = 3 and A as the linear mapping for which

   
µ ¶ 2 µ ¶ 7
1 0
A = 1 , A = 0 ,
0 1
5 3

where components are with respect to the standard bases for R2 and R3 , find the matrix of A with
respect to the bases
     
µ ¶ µ ¶ 1 0 0
1 −3
e1 = , e2 = ; f1 = 0 , f2 = 1 , f3 = 0 .
−1 2
0 0 1

(b) The mapping A of R3 to itself is a reflection in the plane x1 sin θ = x2 cos θ. Find the matrix A of
A with respect to any basis of your choice, which should be specified.

3. The linear map A : R2 → R2 is defined by

µ ¶ µ ′¶ µ ¶ µ ¶
x x x −5x + 9y
7−→ = A = .
y y′ y −4x + 7y

Find the matrix B of the map A relative to the basis

½µ ¶ µ ¶¾
3 1
, ,
2 1

and interpret the map geometrically. Hence show that, for each positive integer n,

Bn − I = n(B − I) ,
µ ¶
1 0
where I = . Hence evaluate An . Verify that det(An ) = (det A)n .
0 1
∗4. Show that similar matrices have the same rank.

5. Show that the matrix  

0 1 0
A = −4 4 0
−2 1 2
has characteristic equation (t − 2)3 = 0. Explain (without doing any further calculations) why A is not
diagonalisable.

Mathematical Tripos IA: Vectors and Matrices 1 [email protected], Michaelmas 2014

 Copies available at http://www.damtp.cam.ac.uk/user/examples/

6. (a) Find a, b and c such that the matrix

 
1/3 0√ a
2/3 1/ √2 b 
2/3 −1/ 2 c

is orthogonal. Does this condition determine a, b and c uniquely?

(b) Let µ ¶ µ ¶
2 0 1 0
A= and P = .
0 1 1 1
Do you expect PAP−1 to be symmetric? Compute PAP−1 . Were you right?
∗7. (a) Show that the Cayley-Hamilton theorem is true for all 2 × 2 matrices.
(b) Let
Copyright © 2014 University of Cambridge. Not to be quoted or reproduced without permission.

µ ¶
3 4
A= .
−1 −1
Find the characteristic equation for A. Verify that A2 = 2A − I. Is A diagonalisable ?
Show by induction that An lies in the two-dimensional subspace (of the space of 2 × 2 real matrices)
spanned by A and I, i.e. show that there exists real numbers αn and βn such that

An = αn A + βn I .

Find a recurrence relation (i.e. a difference equation) for αn and βn , and hence find an explicit
formula for An .
8. Determine the eigenvalues and eigenvectors of the symmetric matrix
 
3 1 1
A = 1 2 0 .
1 0 2

Use an identity of the form PT AP = D, where D is a diagonal matrix, to find A−1 .

∗9. Show that the eigenvalues of a unitary matrix have unit modulus. Show that if a unitary matrix has
distinct eigenvalues then the eigenvectors are orthogonal.
10. A skew-Hermitian matrix, W, is one such that W† = −W. What can be said about the eigenvalues of a
skew-Hermitian matrix? (Hint: consider H = iW)?
If S is real symmetric and T is real skew-symmetric, show that T±iS is skew-Hermitian. State a property
of the eigenvalues of T + iS and hence, or otherwise, show that

det(T + iS − I) 6= 0 .

Show that the matrix

U = (I + T + iS)(I − T − iS)−1
is unitary. For ¶µ µ ¶
1 1 0 1
S= , T= ,
1 1 −1 0
√
show that the eigenvalues of U are ±(1 − i)/ 2.
∗11. This is a continuation of question 8 on Example Sheet 2.
As in question 8 on Example Sheet 2 consider the linear map S : R2 → R2

x 7→ x′ = x + λ(b · x) a (∗)

where λ is a real scalar, a and b are fixed orthogonal unit vectors. Let S(λ, a, b) be the matrix with
elements Sij such that x′i = Sij xj . Give diagrams illustrating the shears

S1 = S(λ, i, j) , S2 = S(λ, j, −i).

Mathematical Tripos IA: Vectors and Matrices 2 [email protected], Michaelmas 2014

 Copies available at http://www.damtp.cam.ac.uk/user/examples/

Show that S1 and S2 are related by a similarity transformation

µ ¶
−1 0 −1
S2 = R S1 R , R = .
1 0

Now let S be the map defined by (∗) but from R3 to R3 , and let i, j, k be unit vectors along the three
perpendicular axes. Find the matrix S in each of the cases

(i) a = i, b = j, (ii) a = j, b = k, (iii) a = k, b = i,

and interpret the corresponding simple shears. Show that any matrix of the form
 
1 λ µ
0 1 ν
0 0 1

can be displayed (not necessarily uniquely) as the product of matrices of simple shears.
Copyright © 2014 University of Cambridge. Not to be quoted or reproduced without permission.

∗12. Diagonalize the quadratic form

F = (a cos2 θ + b sin2 θ)x2 + 2(a − b)(sin θ cos θ)xy + (a sin2 θ + b cos2 θ)y 2 ,

and identify the principal axes.

13. Find all eigenvalues, and an orthonormal set of eigenvectors, of the matrices
 √   
5 0 3 2 −1 −1
A = √0 3 0  and B = −1 2 −1
3 0 3 −1 −1 2.

Hence sketch the surfaces

√
5x2 + 3y 2 + 3z 2 + 2 3xz = 1 and x2 + y 2 + z 2 − xy − yz − zx = 1 .

14. Let Σ be the surface in R3 given by

2x2 + 2xy + 4yz + z 2 = 1.

By writing this equation as

xT Ax = 1,
with A a real symmetric matrix, show that there is an orthonormal basis such that, if we use coordinates
(u, v, w) with respect to this new basis, Σ takes the form

λu2 + µv 2 + νw2 = 1.

Find λ, µ and ν and hence find the minimum distance between the origin and Σ. Hint: it is not
necessary to find the basis explicitly.
15. (i) Explain what is meant by saying that a 2 × 2 real matrix,
µ ¶
a b
A= ,
c d

preserves the scalar product on R2 with respect to

µ ¶ µ ¶
1 0 1 0
(a) the Euclidean metric, I = , or (b) the Minkowski metric, J = .
0 1 0 −1

(ii) Using a single real parameter together with a choice of sign (±1), give and justify a description of
all matrices, A, that preserve the scalar product with respect to the Euclidean metric. Show that
these matrices form a group.
(iii) Using a single real parameter together with a choice of sign (±1), give and justify a description of
all matrices, A with a > 0, that preserve the scalar product with respect to the Minkowski metric.
Show that these matrices form a group.
(iv) What is the intersection of the above two groups?

Mathematical Tripos IA: Vectors and Matrices 3 [email protected], Michaelmas 2014

 Copies available at http://www.damtp.cam.ac.uk/user/examples/

Revision Questions

16. Show that a rotation about the z axis through an angle θ corresponds to the matrix
 
cos θ − sin θ 0
R =  sin θ cos θ 0 .
0 0 1

Write down a real eigenvector of R and give the corresponding eigenvalue. In the case of a matrix
corresponding to a general rotation, how can one find the axis of rotation?
A rotation through 45◦ about the x-axis is followed by a similar one about the z-axis. Show that the
rotation corresponding to their combined effect has its axis inclined at equal angles
1
cos−1 √ √
(5 − 2 2)
Copyright © 2014 University of Cambridge. Not to be quoted or reproduced without permission.

to the x and z axes.

17. Show that the eigenvalues of a real orthogonal matrix have unit modulus and that if λ is an eigenvalue
so is λ∗ . Hence argue that the eigenvalues of a 3 × 3 real orthogonal matrix Q must be a selection from

+1 , −1 and eiα & e−iα .

Verify that det Q = ±1. What is the effect of Q on vectors orthogonal to an eigenvector with eigen-
value ±1?
∗18. This is another way of proving det AB = det A det B. You may wish to stick to the case n = 3.
If 1 6 r, s 6 n, r 6= s and λ is real, let E(λ, r, s) be an n × n matrix with (i, j) entry δij + λδir δjs . If
1 6 r 6 n and µ is real, let F (µ, r) be an n × n matrix with (i, j) entry δij + (µ − 1)δir δjr .

(i) Give a simple geometric interpretation of the linear maps from Rn to Rn associated with E(λ, r, s)
and F(µ, r).
(ii) Give a simple account of the effect of pre-multiplying an n × m matrix by E(λ, r, s) and by F(µ, r).
What is the effect of post-multiplying an m × n matrix?
(iii) If A is an n × n matrix, find det(E(λ, r, s)A) and det(F(µ, r)A) in terms of det A.
(iv) Show that every n × n matrix is the product of matrices of the form E(λ, r, s) and F(µ, r) and a
diagonal matrix with entries 0 or 1.
(v) Use (iii) and (iv) to show that, if A and B are n × n matrices, then det A det B = det AB.

∗19. The object of this exercise is to show why finding eigenvalues of a large matrix is not just a matter of
finding a large fast computer.
Consider the n × n complex matrix A = {Aij } given by

Aj j+1 = 1 for 1 6 j 6 n − 1
An1 = κn
Aij = 0 otherwise,

where κ ∈ C is non-zero. Thus, when n = 2 and n = 3, we get the matrices

 
µ ¶ 0 1 0
0 1
and  0 0 1 .
κ2 0
κ3 0 0

(i) Find the eigenvalues and associated eigenvectors of A for n = 2 and n = 3.

(ii) By guessing and then verifying your answers, or otherwise, find the eigenvalues and associated
eigenvectors of A for for all n > 2.
(iii) Suppose that your computer works to 15 decimal places and that n = 100. You decide to find the
eigenvalues of A in the cases κ = 2−1 and κ = 3−1 . Explain why at least one (and more probably)
both attempts will deliver answers which bear no relation to the true answers.

Mathematical Tripos IA: Vectors and Matrices 4 [email protected], Michaelmas 2014