MATHEMATICAL TRIPOS Part II

Friday, 7 June, 2019 9:00 am to 12:00 pm MAT2

PAPER 4

Before you begin read these instructions carefully.

The examination paper is divided into two sections. Each question in Section II
carries twice the number of marks of each question in Section I. Section II questions
also carry an alpha or beta quality mark and Section I questions carry a beta quality
mark.
Candidates may obtain credit from attempts on at most six questions from Section
I and from any number of questions from Section II.
Write on one side of the paper only and begin each answer on a separate sheet.
Write legibly; otherwise you place yourself at a grave disadvantage.

At the end of the examination:

Tie up your answers in bundles, marked A, B, C, . . ., J according to the code letter
affixed to each question. Include in the same bundle all questions from Sections I
and II with the same code letter.
Attach a completed gold cover sheet to each bundle.
You must also complete a green master cover sheet listing all the questions you have
attempted.
Every cover sheet must bear your examination number and desk number.

STATIONERY REQUIREMENTS
Gold cover sheet
Green master cover sheet
Script paper
Rough paper

You may not start to read the questions

printed on the subsequent pages until
instructed to do so by the Invigilator.
 2

SECTION I
1I Number Theory
Show that the product
Y  1
−1
1−
p
p prime

and the series X 1

p
p prime

are both divergent.

2H Topics in Analysis
Show that π is irrational. [Hint: consider the functions fn : [0, π] → R given by
fn (x) = xn (π − x)n sin x.]

3G Coding and Cryptography

(a) Describe Diffie-Hellman key exchange. Why is it believed to be a secure system?
(b) Consider the following authentication procedure. Alice chooses public key N
for the Rabin–Williams cryptosystem. To be sure we are in communication with Alice we
send her a ‘random item’ r ≡ m2 mod N . On receiving r, Alice proceeds to decode using
her knowledge of the factorisation of N and finds a square root m1 of r. She returns m1
to us and we check r ≡ m21 mod N . Is this authentication procedure secure? Justify your
answer.

4H Automata and Formal Languages

(a) Which of the following are regular languages? Justify your answers.

(i) {wn | w ∈ {a, b}∗ , n > 2}.

(ii) {w ∈ {a, b, c}∗ | w contains an odd number of b’s and an even number of c’s}.

(iii) {w ∈ {0, 1}∗ | w contains no more than 7 consecutive 0’s}.

(b) Consider the language L over alphabet {a, b} defined via

L := {wabn | w ∈ {a, b}∗ , n ∈ K} ∪ {b}∗ .

Show that L satisfies the pumping lemma for regular languages but is not a regular
language itself.

Part II, Paper 4

 3

5J Statistical Modelling
In a normal linear model with design matrix X ∈ Rn×p , output variables y ∈ Rn
and parameters β ∈ Rp and σ 2 > 0, define a (1 − α)-level prediction interval for a new
observation with input variables x∗ ∈ Rp . Derive an explicit formula for the interval,
proving that it satisfies the properties required by the definition. [You may assume that
the maximum likelihood estimator β̂ is independent of σ −2 ky − X β̂k22 , which has a χ2n−p
distribution.]

6C Mathematical Biology
(a) A variant of the classic logistic population model is given by:

dx(t)  
= α x(t) − x(t − T )2
dt
where α, T > 0.
Show that for small T , the constant solution x(t) = 1 is stable.
Allow T to increase. Express in terms of α the value of T at which the constant
solution x(t) = 1 loses stability.
(b) Another variant of the logistic model is given by this equation:

dx(t)
= α x(t − T ) [1 − x(t)]
dt
where α, T > 0. When is the constant solution x(t) = 1 stable for this model?

7A Further Complex Methods

A single-valued function Arcsin(z) can be defined, for 0 6 arg z < 2π, by means of
an integral as:
Z z
dt
Arcsin(z) = . (†)
0 (1 − t2 )1/2

(a) Choose a suitable branch-cut with the integrand taking a value +1 at the origin
on the upper side of the cut, i.e. at t = 0+ , and describe suitable paths of integration in
the two cases 0 6 arg z 6 π and π < arg z < 2π.
(b) Construct the multivalued function arcsin(z) by analytic continuation.


(c) Express arcsin e2πi z in terms of Arcsin(z) and deduce the periodicity property
of sin(z).

Part II, Paper 4 [TURN OVER]

 4

8E Classical Dynamics
(a) The angular momentum of a rigid body about its centre of mass is conserved.
Derive Euler’s equations,

I1 ω̇1 = (I2 − I3 )ω2 ω3 ,

I2 ω̇2 = (I3 − I1 )ω3 ω1 ,
I3 ω̇3 = (I1 − I2 )ω1 ω2 ,

explaining the meaning of the quantities appearing in the equations.

(b) Show that there are two independent conserved quantities that are quadratic
functions of ω = (ω1 , ω2 , ω3 ), and give a physical interpretation of them.
(c) Derive a linear approximation to Euler’s equations that applies when |ω1 | ≪ |ω3 |
and |ω2 | ≪ |ω3 |. Use this to determine the stability of rotation about each of the three
principal axes of an asymmetric top.

9B Cosmology
Derive the relation between the neutrino temperature Tν and the photon tempera-
ture Tγ at a time long after electrons and positrons have become non-relativistic.
[In this question you may work in units of the speed of light, so that c = 1. You may
also use without derivation the following formulae. The energy density ǫa and pressure Pa
for a single relativistic species a with a number ga of degenerate states at temperature T
are given by
Z Z
4πga p3 dp 4πga p3 dp
ǫa = 3 , Pa = ,
h ep/(kB T ) ∓ 1 3h3 ep/(kB T ) ∓ 1
where kB is Boltzmann’s constant, h is Planck’s constant, and the minus or plus depends
on whether the particle is a boson or a fermion respectively. For each species a, the entropy
density sa at temperature Ta is given by,
ǫa + Pa
sa = .
kB Ta
The effective total number g∗ of relativistic species is defined in terms of the numbers of
bosonic and fermionic particles in the theory as,
X 7 X
g∗ = gbosons + gf ermions ,
8
bosons f ermions

with the specific values gγ = ge+ = ge− = 2 for photons, positrons and electrons.]

Part II, Paper 4

 5

10D Quantum Information and Computation

(a) Define the order of α mod N for coprime integers α and N with α < N . Explain
briefly how knowledge of this order can be used to provide a factor of N , stating conditions
on α and its order that must be satisfied.
(b) Shor’s algorithm for factoring N starts by choosing α < N coprime. Describe
the subsequent steps of a single run of Shor’s algorithm that computes the order of α mod
N with probability O(1/ log log N ).
[Any significant theorems that you invoke to justify the algorithm should be clearly
stated (but proofs are not required). In addition you may use without proof the following
two technical results.
Theorem FT: For positive integers t and M with M > t2 , and any 0 6 x0 < t, let K
be the largest integer such that x0 + (K − 1)t <M . Let QF T denote the quantum Fourier
P
transform mod M . Suppose we measure QF T √1K K−1 k=0 |x0 + kti to obtain an integer
c with 0 6 c < M . Then with probability O(1/ log log t), c will be an integer closest to a
multiple j(M/t) of M/t for which the value of j (between 0 and t) is coprime to t.
Theorem CF: For any rational number a/b with 0 < a/b < 1 and with integers a
and b having at most n digits each, let p/q with p and q coprime, be any rational number
satisfying  
a p
 − 6 1 .
 b q  2q 2

Then p/q is one of the O(n) convergents of the continued fraction of a/b and all the
convergents can be classically computed from a/b in time O(n3 ).]

Part II, Paper 4 [TURN OVER]

 6

SECTION II
11I Number Theory
(a) Let a0 , a1 , . . . be positive integers, and β > 0 a positive real number. Show that
for every n > 0, if θn = [a0 , . . . , an , β], then θn = (βpn + pn−1 )/(βqn + qn−1 ), where (pn ),
(qn ) (n > −1) are sequences of integers satisfying

p0 = a0 , q0 = 1, p−1 = 1, q−1 = 0 and

    
pn pn−1 pn−1 pn−2 an 1
= (n > 1).
qn qn−1 qn−1 qn−2 1 0

Show that pn qn−1 − pn−1 qn = (−1)n−1 , and that θn lies between pn /qn and pn−1 /qn−1 .
(b) Show that if [a0 , a1 , . . . ] is the continued fraction expansion of a positive
irrational θ, then pn /qn → θ as n → ∞.
(c) Let the convergents of the continued fraction [a0 , a1 , . . . , an ] be pj /qj (0 6
j 6 n). Using part (a) or otherwise, show that the n-th and (n − 1)-th convergents
of [an , an−1 , . . . , a0 ] are pn /pn−1 and qn /qn−1 respectively.
(d) Show that if θ = [ a0 , a1 , . . . , an ] is a purely periodic continued fraction with
convergents pj /qj , then f (θ) = 0, where f (X) = qn X 2 + (qn−1 − pn )X − pn−1 . Deduce
that if θ ′ is the other root of f (X), then −1/θ ′ = [ an , an−1 , . . . , a0 ].

12H Topics in Analysis

(a) Suppose that K ⊂ C is a non-empty subset of the square {x+ iy : x, y ∈ (−1, 1)}
and f is analytic in the larger square {x + iy : x, y ∈ (−1 − δ, 1 + δ)} for some δ > 0. Show
that f can be uniformly approximated on K by polynomials.
(b) Let K be a closed non-empty proper subset of C. Let Λ be the set of λ ∈ C \ K
such that gλ (z) = (z − λ)−1 can be approximated uniformly on K by polynomials and let
Γ = C \ (K ∪ Λ). Show that Λ and Γ are open. Is it always true that Λ is non-empty? Is
it always true that, if K is bounded, then Γ is empty? Give reasons.
[No form of Runge’s theorem may be used without proof.]

Part II, Paper 4

 7

13J Statistical Modelling

A sociologist collects a dataset on friendships among m Cambridge graduates. Let
yi,j = 1 if persons i and j are friends 3 years after graduation, and yi,j = 0 otherwise. Let
zi be a categorical variable for person i’s college, taking values in the set {1, 2, . . . , C}.
Consider logistic regression models,

eθi,j
P(yi,j = 1) = , 1 6 i < j 6 m,
1 + eθi,j
with parameters either

1. θi,j = βzi ,zj ; or,

2. θi,j = βzi + βzj ; or,

3. θi,j = βzi + βzj + β0 δzi ,zj , where δzi ,zj = 1 if zi = zj and 0 otherwise.

(a) Write the likelihood of the models.

(b) Show that the three models are nested and specify the order. Suggest a statistic
to compare models 1 and 3, give its definition and specify its asymptotic distribution under
the null hypothesis, citing any necessary theorems.
(c) Suppose persons i and j are in the same college k; consider the number of
friendships, Mi and Mj , that each of them has with people in college ℓ 6= k (ℓ and
k fixed). In each of the models above, compare the distribution of these two random
variables. Explain why this might lead to a poor quality of fit.
(d) Find a minimal sufficient statistic for model 3. [You may use the following
characterisation of a minimal sufficient statistic: let f (β; y) be the likelihood in this model,
where β = (βk )k=0,1,...,C and y = (yi,j )i,j=1,...,m ; suppose T = t(y) is a statistic such that
f (β; y)/f (β; y ′ ) is constant in β if and only if t(y) = t(y ′ ); then, T is a minimal sufficient
statistic for β.]

Part II, Paper 4 [TURN OVER]

 8

14C Mathematical Biology

A model of an infectious disease in a plant population is given by

Ṡ = (S + I) − (S + I)S − βIS , (1)

I˙ = −(S + I)I + βIS (2)

where S(t) is the density of healthy plants and I(t) is the density of diseased plants at
time t and β is a positive constant.
(a) Give an interpretation of what each of the terms in equations (1) and (2)
represents in terms of the dynamics of the plants. What does the coefficient β represent?
What can you deduce from the equations about the effect of the disease on the plants?
(b) By finding all fixed points for S > 0 and I > 0 and analysing their stability,
explain what will happen to a healthy plant population if the disease is introduced. Sketch
the phase diagram, treating the cases β < 1 and β > 1 separately.
(c) Define new variables N (t) for the total plant population density and θ(t) for the
proportion of the population that is diseased. Starting from equations (1) and (2) above,
derive equations for Ṅ and θ̇ purely in terms of N , θ and β. Without carrying out a full
fixed point analysis, explain how this system can be used directly to show the same results
you had in part (b). [Hint: start by considering the dynamics of N (t) alone.]
(d) Suppose now that in an attempt to control disease, plants are culled at a rate k
per capita, independently of whether the plants are healthy or diseased. Write down the
modified versions of equations (1) and (2). Use these to build updated equations for Ṅ and
θ̇. Without carrying out a detailed fixed point analysis, what can you deduce about the
effect of culling? Give the range of k for which culling can effectively control the disease.

Part II, Paper 4

 9

15E Classical Dynamics

(a) Explain what is meant by a Lagrange top. You may assume that such a top has
the Lagrangian
1   1  2
L = I1 θ̇ 2 + φ̇2 sin2 θ + I3 ψ̇ + φ̇ cos θ − M gl cos θ
2 2
in terms of the Euler angles (θ, φ, ψ). State the meaning of the quantities I1 , I3 , M and l
appearing in this expression.
Explain why the quantity
∂L
pψ =
∂ ψ̇
is conserved, and give two other independent integrals of motion.
Show that steady precession, with a constant value of θ ∈ (0, π2 ), is possible if

p2ψ > 4M glI1 cos θ .

(b) A rigid body of mass M is of uniform density and its surface is defined by

x33
x21 + x22 = x23 − ,
h
where h is a positive constant and (x1 , x2 , x3 ) are Cartesian coordinates in the body frame.
Calculate the values of I1 , I3 and l for this symmetric top, when it rotates about
the sharp point at the origin of this coordinate system.

16I Logic and Set Theory

Define the cardinals ℵα , and explain briefly why every infinite set has cardinality
an ℵα .
Show that if κ is an infinite cardinal then κ2 = κ.
Let X1 , X2 , . . . , Xn be infinite sets. Show that X1 ∪ X2 ∪ · · · ∪ Xn must have the
same cardinality as Xi for some i.
Let X1 , X2 , . . . be infinite sets, no two of the same cardinality. Is it possible that
X1 ∪ X2 ∪ . . . has the same cardinality as some Xi ? Justify your answer.

Part II, Paper 4 [TURN OVER]

 10

17G Graph Theory

State and prove Hall’s theorem.
Let n be an even positive integer. Let X = {A : A ⊂ [n]} be the power set of
[n] = {1, 2, . . . , n}. For 1 6 i 6 n, let Xi = {A ∈ X : |A| = i}. Let Q be the graph with
vertex set X where A, B ∈ X are adjacent if and only if |A△B| = 1. [Here, A△B denotes
the symmetric difference of A and B, given by A△B := (A ∪ B) \ (A ∩ B).]
Let 1 6 i 6 n2 . Why is the induced subgraph Q[Xi ∪ Xi−1 ] bipartite? Show that it
contains a matching from Xi−1 to Xi .
A chain in X is a subset C ⊂ X such that whenever A, B ∈ C we have A ⊂ B or
B ⊂ A. What is the least positive integer k such that X can be partitioned into k pairwise
disjoint chains? Justify your answer.

18F Galois Theory

State (without proof) a result concerning uniqueness of splitting fields of a polyno-
mial.
Given a polynomial f ∈ Q[X] with distinct roots, what is meant by its Galois group
GalQ (f )? Show that f is irreducible over Q if and only if GalQ (f ) acts transitively on the
roots of f .
Now consider an irreducible quartic of the form g(X) = X 4 + bX 2 + c ∈ Q[X]. If
√
α ∈ C denotes a root of g, show that the splitting field K ⊂ C is Q(α, c). Give an
explicit description of Gal(K/Q) in the cases:
√
(i) c ∈ Q(α), and
√
(ii) c 6∈ Q(α).

If c is a square in Q, deduce that GalQ (g) ∼= C2 × C2 . Conversely, if GalQ (g) ∼

=
√
C2 × C2 , show that c is invariant under at least two elements of order two in the Galois
group, and deduce that c is a square in Q.

Part II, Paper 4

 11

19I Representation Theory

(a) What is meant by a compact topological group? Explain why SU(n) is an example
of such a group.
[In the following the existence of a Haar measure for any compact Hausdorff
topological group may be assumed, if required.]
(b) Let G be any compact Hausdorff topological group. Show that there is a
continuous group homomorphism ρ : G → O(n) if and only if G has an n-dimensional
representation over R. [Here O(n) denotes the subgroup of GLn (R) preserving the standard
(positive-definite) symmetric bilinear form.]
(c) Explicitly construct such a representation ρ : SU(2) → SO(3) by showing that
SU(2) acts on the following vector space of matrices,
   
a b t
A= ∈ M2 (C) : A + A = 0
c −a

by conjugation.
Show that

(i) this subspace is isomorphic to R3 ;

(ii) the trace map (A, B) 7→ −tr(AB) induces an invariant positive definite
symmetric bilinear form;

(iii) ρ is surjective with kernel {±I2 }. [You may assume, without proof, that SU(2)
is connected.]

Part II, Paper 4 [TURN OVER]

 12

20G Number Fields

(a) Let L be a number field, and suppose there exists α ∈ OL such that OL = Z[α].
Let f (X) ∈ Z[X] denote the minimal polynomial of α, and let p be a prime. Let
f (X) ∈ (Z/pZ)[X] denote the reduction modulo p of f (X), and let

f (X) = g 1 (X)e1 · · · gr (X)er

denote the factorisation of f (X) in (Z/pZ)[X] as a product of powers of distinct monic

irreducible polynomials g1 (X), . . . , g r (X), where e1 , . . . , er are all positive integers.
For each i = 1, . . . , r, let gi (X) ∈ Z[X] be any polynomial with reduction modulo p
equal to g i (X), and let Pi = (p, gi (α)) ⊂ OL . Show that P1 , . . . , Pr are distinct, non-zero
prime ideals of OL , and that there is a factorisation

pOL = P1e1 · · · Prer ,

and that N (Pi ) = pdeg gi (X) .

(b) Let K be a number field of degree n = [K : Q], and let p be a prime. Suppose
that there is a factorisation
pOK = Q1 · · · Qs ,
where Q1 , . . . , Qs are distinct, non-zero prime ideals of OK with N (Qi ) = p for each i =
1, . . . , s. Use the result of part (a) to show that if n > p then there is no α ∈ OK such that
OK = Z[α].

21F Algebraic Topology

State the Lefschetz fixed point theorem.
Let n > 2 be an integer, and x0 ∈ S 2 a choice of base point. Define a space

X := (S 2 × Z/nZ)/ ∼

where Z/nZ is discrete and ∼ is the smallest equivalence relation such that (x0 , i) ∼
(−x0 , i + 1) for all i ∈ Z/nZ. Let φ : X → X be a homeomorphism without fixed points.
Use the Lefschetz fixed point theorem to prove the following facts.

(i) If φ3 = IdX then n is divisible by 3.

(ii) If φ2 = IdX then n is even.

Part II, Paper 4

 13

22H Linear Analysis

(a) State and prove the Riesz representation theorem for a real Hilbert space H.
[You may use that if H is a real Hilbert space and Y ⊂ H is a closed subspace, then
H = Y ⊕ Y ⊥ .]
(b) Let H be a real Hilbert space and T : H → H a bounded linear operator. Show
that T is invertible if and only if both T and T ∗ are bounded below. [Recall that an
operator S : H → H is bounded below if there is c > 0 such that kSxk > ckxk for all
x ∈ H.]
(c) Consider the complex Hilbert space of two-sided sequences,
X
X = {(xn )n∈Z : xn ∈ C, |xn |2 < ∞}
n∈Z
P
with norm kxk = ( n |xn |2 )1/2 . Define T : X → X by (T x)n = xn+1 . Show that T is
unitary and find the point spectrum and the approximate point spectrum of T .

23H Analysis of Functions

(a) Let (H, h·, ·i) be a real Hilbert space and let B : H × H → R be a bilinear map.
If B is continuous prove that there is an M > 0 such that |B(u, v)| 6 M kukkvk for all
u, v ∈ H. [You may use any form of the Banach–Steinhaus theorem as long as you state
it clearly.]
(b) Now suppose that B defined as above is bilinear and continuous, and assume
also that it is coercive: i.e. there is a C > 0 such that B(u, u) > Ckuk2 for all u ∈ H.
Prove that for any f ∈ H, there exists a unique vf ∈ H such that B(u, vf ) = hu, f i for all
u ∈ H.
[Hint: show that there is a bounded invertible linear operator L with bounded
inverse so that B(u, v) = hu, Lvi for all u, v ∈ H. You may use any form of the Riesz
representation theorem as long as you state it clearly.]
(c) Define the Sobolev space H01 (Ω), where Ω ⊂ Rd is open and bounded.
(d) Suppose f ∈ L2 (Ω) and A ∈ Rd with |A|2 < 2, where | · |2 is the Euclidean norm
on Rd .Consider the Dirichlet problem

−∆v + v + A · ∇v = f in Ω, v=0 in ∂Ω.

Using the result of part (b), prove there is a unique weak solution v ∈ H01 (Ω).
(e) Now assume that Ω is the open unit disk in R2 and g is a smooth function on
S1 . Sketch how you would solve the following variant:

−∆v + v + A · ∇v = 0 in Ω, v=g in ∂Ω.

[Hint: Reduce to the result of part (d).]

Part II, Paper 4 [TURN OVER]

 14

24F Algebraic Geometry

(a) Let X ⊆ P2 be a smooth projective plane curve, defined by a homogeneous
polynomial F (x, y, z) of degree d over the complex numbers C.

(i) Define the divisor [X ∩ H], where H is a hyperplane in P2 not contained in

X, and prove that it has degree d.

(ii) Give (without proof) an expression for the degree of KX in terms of d.

(iii) Show that X does not have genus 2.

(b) Let X be a smooth projective curve of genus g over the complex numbers C.
For p ∈ X let

G(p) = {n ∈ N | there is no f ∈ k(X) with vp (f ) = n, and vq (f ) 6 0 for all q 6= p}.

(i) Define ℓ(D), for a divisor D.

(ii) Show that for all p ∈ X,

(
ℓ((n − 1)p) for n ∈ G(p)
ℓ(np) =
ℓ((n − 1)p) + 1 otherwise.

(iii) Show that G(p) has exactly g elements. [Hint: What happens for large n? ]

(iv) Now suppose that X has genus 2. Show that G(p) = {1, 2} or G(p) = {1, 3}.

[In this question N denotes the set of positive integers.]

Part II, Paper 4

 15

25H Differential Geometry

(a) Let γ : (a, b) → R2 be a regular curve without self-intersection given by
γ(v) = (f (v), g(v)) with f (v) > 0 for v ∈ (a, b) and let S be the surface of revolution
defined globally by the parametrisation

φ : (0, 2π) × (a, b) → R3 ,

where φ(u, v) = (f (v) cos u, f (v) sin u, g(v)), i.e. S = φ((0, 2π) × (a, b)). Compute its mean
curvature H and its Gaussian curvature K.
(b) Define what it means for a regular surface S ⊂ R3 to be minimal. Give an
example of a minimal surface which is not locally isometric to a cone, cylinder or plane.
Justify your answer.
(c) Let S be a regular surface such that K ≡ 1. Is it necessarily the case that given
any p ∈ S, there exists an open neighbourhood U ⊂ S of p such that U lies on some sphere
in R3 ? Justify your answer.

Part II, Paper 4 [TURN OVER]

 16

26K Probability and Measure

(a) Let (Xn )n>1 and X be real random variables with finite second moment on a
probability space (Ω, F, P). Assume that Xn converges to X almost surely. Show that the
following assertions are equivalent:

(i) Xn → X in L2 as n → ∞,

(ii) E(Xn2 ) → E(X 2 ) as n → ∞.

(b) Suppose now that Ω = (0, 1), F is the Borel σ-algebra of (0, 1) and P is Lebesgue
measure. Given a Borel probability measure µ on R we set

Xµ (ω) = inf{x ∈ R|Fµ (x) > ω},

where Fµ (x) := µ((−∞, x]) is the distribution function of µ and ω ∈ Ω.

(i) Show that Xµ is a random variable on (Ω, F, P) with law µ.

(ii) Let (µn )n>1 and ν be Borel probability measures on R with finite second
moments. Show that

E((Xµn − Xν )2 ) → 0 as n → ∞
R
Rif and only if µn converges weakly to ν and x2 dµn (x) converges to
x2 dν(x) as n → ∞.

[You may use any theorem proven in lectures as long as it is clearly stated.
Furthermore, you may use without proof the fact that µn converges weakly to ν as n → ∞
if and only if Xµn converges to Xν almost surely.]

Part II, Paper 4

 17

27K Applied Probability R

(a) Let λ : Rd → [0, ∞) be such that Λ(A) := A λ(x) dx is finite for any bounded
measurable set A ⊆ Rd . State the properties which define a (non-homogeneous) Poisson
process Π on Rd with intensity function λ.
(b) Let Π be a Poisson process on Rd with intensity function λ, and let f : Rd → Rs
be a given function. Give a clear statement of the necessary conditions on the pair Λ, f
subject to which f (Π) is a Poisson process on Rs . When these conditions hold, express
the mean measure of f (Π) in terms of Λ and f .
(c) Let Π be a homogeneous Poisson process on R2 with constant intensity 1, and
let f : R2 → [0, ∞) be given by f (x1 , x2 ) = x21 + x22 . Show that f (Π) is a homogeneous
Poisson process on [0, ∞) with constant intensity π.
Let R1 , R2 , . . . be an increasing sequence of positive random variables such that the
points of f (Π) are R12 , R22 , . . . . Show that Rk has density function

1 2
hk (r) = 2πr(πr 2 )k−1 e−πr , r > 0.
(k − 1)!

Part II, Paper 4 [TURN OVER]

 18

28J Principles of Statistics

We consider a statistical model {f (·, θ) : θ ∈ Θ}.
(a) Define the maximum likelihood estimator (MLE) and the Fisher information
I(θ).
(b) Let Θ = R and assume there exist a continuous one-to-one function µ : R → R
and a real-valued function h such that

Eθ [h(X)] = µ(θ) ∀θ ∈ R.

(i) For X1 , . . . , Xn i.i.d. from the model for some θ0 ∈ R, give the limit in almost
sure sense of
n
1X
µ̂n = h(Xi ) .
n
i=1

Give a consistent estimator θ̂n of θ0 in terms of µ̂n .

(ii) Assume further that Eθ0 [h(X)2 ] < ∞ and that µ is continuously differentiable
√
and strictly monotone. What is the limit in distribution of n(θ̂n − θ0 )? Assume too that
the statistical model satisfies the usual regularity assumptions. Do you necessarily expect
Var(θ̂n ) > (nI(θ0 ))−1 for all n? Why?
(iii) Propose an alternative estimator for θ0 with smaller bias than θ̂n if Bn (θ0 ) =
Eθ0 [θ̂n ] − θ0 = na + nb2 + O( n13 ) for some a, b ∈ R with a 6= 0.
(iv) Further to all the assumptions in iii), assume that the MLE for θ0 is of the form
n
1X
θ̂M LE = h(Xi ).
n
i=1

What is the link between the Fisher information at θ0 and the variance of h(X)? What
does this mean in terms of the precision of the estimator and why?
[You may use results from the course, provided you state them clearly.]

Part II, Paper 4

 19

29K Stochastic Financial Models

(a) Describe the (Cox-Ross-Rubinstein) binomial model. What are the necessary
and sufficient conditions on the model parameters for it to be arbitrage-free? How is the
equivalent martingale measure Q characterised in this case?
(b) Consider a discounted claim H of the form H = h(S01 , S11 , . . . , ST1 ) for some
function h. Show that the value process of H is of the form


Vt (ω) = vt S01 , S11 (ω), . . . , St1 (ω) ,

for t ∈ {0, . . . , T }, where the function vt is given by

" #
 S11 ST1 −t 
vt (x0 , . . . , xt ) = EQ h x0 , . . . , xt , xt · 1 , . . . , xt · .
S0 S01

You may use any property of conditional expectations without proof.

(c) Suppose that H = h(ST1 ) only depends on the terminal value ST1 of the stock
price. Derive an explicit formula for the value of H at time t ∈ {0, . . . , T }.
(d) Suppose that H is of the form H = h(ST1 , MT ), where Mt := maxs∈{0,...,t} Ss1 .
Show that the value process of H is of the form


Vt (ω) = vt St1 (ω), Mt (ω) ,

for t ∈ {0, . . . , T }, where the function vt is given by

 
vt (x, m) = EQ g(x, m, S01 , ST1 −t , MT −t )

for a function g to be determined.

Part II, Paper 4 [TURN OVER]

 20

30A Asymptotic Methods

Consider, for small ǫ, the equation

d2 ψ
ǫ2 − q(x)ψ = 0. (∗)
dx2
Assume that (∗) has bounded solutions with two turning points a, b where b > a, q ′ (b) > 0
and q ′ (a) < 0.
(a) Use the WKB approximation to derive the relationship
Z b  
1 1/2 1
|q(ξ)| dξ = n+ π with n = 0, 1, 2, · · · . (∗∗)
ǫ a 2

[You may quote without proof any standard results or formulae from WKB theory.]
(b) In suitable units, the radial Schrödinger equation for a spherically symmetric
potential given by V (r) = −V0 /r, for constant V0 , can be recast in the standard form (∗)
as: "   #
~2 d2 ψ 2x x ~2 1 2 −2x
+e λ − V (e ) − l+ e ψ = 0,
2m dx2 2m 2
√
where r = ex and ǫ = ~/ 2m is a small parameter.
Use result (∗∗) to show that the energies of the bound states (i.e λ = −|λ| < 0) are
approximated by the expression:

m V02
E = −|λ| = − .
2~2 (n + l + 1)2

[You may use the result

Z h√
b
1p √ i2
(r − a)(b − r) dr = (π/2) b − a . ]
a r

Part II, Paper 4

 21

31E Dynamical Systems

Consider the dynamical system

ẋ = x + y 2 − a ,
ẏ = y(4x − x2 − a) ,

for (x, y) ∈ R2 , a ∈ R.
Find all fixed points of this system. Find the three different values of a at which
bifurcations appear. For each such value give the location (x, y) of all bifurcations. For
each of these, what types of bifurcation are suggested from this analysis?
Use centre manifold theory to analyse these bifurcations. In particular, for each
bifurcation derive an equation for the dynamics on the extended centre manifold and
hence classify the bifurcation.

32B Principles of Quantum Mechanics

Define the spin raising and spin lowering operators S+ and S− . Show that
p
S± |s, σi = ~ s(s + 1) − σ(σ ± 1) |s, σ ± 1i ,

where Sz |s, σi = ~σ|s, σi and S 2 |s, σi = s(s + 1)~2 |s, σi.

Two spin- 12 particles, with spin operators S(1) and S(2) , have a Hamiltonian

H = αS(1) · S(2) + B · (S(1) − S(2) ) ,

where α and B = (0, 0, B) are constants. Express H in terms of the two particles’ spin
(1) (2) (1)
raising and spin lowering operators S± , S± and the corresponding z-components Sz ,
(2)
Sz . Hence find the eigenvalues of H. Show that there is a unique groundstate in the
limit B → 0 and that the first excited state is triply degenerate in this limit. Explain
this degeneracy by considering the action of the combined spin operator S(1) + S(2) on the
energy eigenstates.

Part II, Paper 4 [TURN OVER]

 22

33B Applications of Quantum Mechanics

(a) A classical beam of particles scatters off a spherically symmetric potential V (r).
Draw a diagram to illustrate the differential cross-section dσ/dΩ, and use this to derive
an expression for dσ/dΩ in terms of the impact parameter b and the scattering angle θ.
A quantum beam of particles of mass m and momentum p = ~k is incident along the
z-axis and scatters off a spherically symmetric potential V (r). Write down the asymptotic
form of the wavefunction ψ in terms of the scattering amplitude f (θ). By considering
the probability current J = −i(~/2m) (ψ ⋆ ∇ψ − (∇ψ ⋆ )ψ), derive an expression for the
differential cross-section dσ/dΩ in terms of f (θ).
(b) The solution ψ(r) of the radial Schrödinger equation for a particle of mass m
and wave number k moving in a spherically symmetric potential V (r) has the asymptotic
form
∞
"

#
X sin kr − lπ cos kr − lπ
2 2
ψ(r) ∼ Al (k) − Bl (k) Pl (cos θ) ,
kr kr
l=0

valid for kr ≫ 1, where Al (k) and Bl (k) are constants and Pl denotes the l’th Legendre
polynomial. Define the S-matrix element Sl and the corresponding phase shift δl for
the partial wave of angular momentum l, in terms of Al (k) and Bl (k). Define also the
scattering length as for the potential V .
Outside some core region, r > r0 , the Schrödinger equation for some such potential
is solved by the s-wave (i.e. l = 0) wavefunction ψ(r) = ψ(r) with,

e−ikr k + iλ tanh(λr) eikr

ψ(r) = +
r k − iλ r
where λ > 0 is a constant. Extract the S-matrix element S0 , the phase shift δ0 and the
scattering length as . Deduce that the potential V (r) has a bound state of zero angular
momentum and compute its energy. Give the form of the (un-normalised) bound state
wavefunction in the region r > r0 .

Part II, Paper 4

 23

34D Statistical Physics

Give an outline of the Landau theory of phase transitions for a system with one real
order parameter φ. Describe the phase transitions that can be modelled by the Landau
potentials

1 1
(i) G = φ4 + εφ2 ,
4 2

1 1 1
(ii) G = φ6 + gφ4 + εφ2 ,
6 4 2

where ε and g are control parameters that depend on the temperature and pressure.
In case (ii), find the curve of first-order phase transitions in the (g, ε) plane. Find
the region where it is possible for superheating to occur. Find also the region where it is
possible for supercooling to occur.

Part II, Paper 4 [TURN OVER]

 24

35E Electrodynamics
Consider a medium in which the electric displacement D(t, x) and magnetising
field H(t, x) are linearly related to the electric and magnetic fields respectively with
corresponding polarisation constants ε and µ;

D = ε E, B = µ H.

Write down Maxwell’s equations for E, B, D and H in the absence of free charges and
currents.
Consider EM waves of the form,

E(t, x) = E0 sin (k · x − ωt) ,

B(t, x) = B0 sin (k · x − ωt) .

Find conditions on the electric and magnetic polarisation vectors E0 and B0 , wave-vector k
and angular frequency ω such that these fields satisfy Maxwell’s equations for the medium
described above. At what speed do the waves propagate?
Consider two media, filling the regions x < 0 and x > 0 in three dimensional space,
and having two different values ε− and ε+ of the electric polarisation constant. Suppose
an electromagnetic wave is incident from the region x < 0 resulting in a transmitted wave
in the region x > 0 and also a reflected wave for x < 0. The angles of incidence, reflection
and transmission are denoted θI , θR and θT respectively. By constructing a corresponding
solution of Maxwell’s equations, derive the law of reflection θI = θR and Snell’s law of
√
refraction, n− sin θI = n+ sin θT where n± = c ε± µ are the indices of refraction of the
two media.
Consider the special case in which the electric polarisation vectors EI , ER and ET of
the incident, reflected and transmitted waves are all normal to the plane of incidence (i.e.
the plane containing the corresponding wave-vectors). By imposing appropriate boundary
conditions for E and H at x = 0, show that,
 
|ER | 1 tan θR
= 1− .
|ET | 2 tan θT

Part II, Paper 4

 25

36D General Relativity

(a) Consider the spherically symmetric spacetime metric

ds2 = −λ2 dt2 + µ2 dr 2 + r 2 dθ 2 + r 2 sin2 θ dφ2 , (†)

where λ and µ are functions of t and r. Use the Euler-Lagrange equations for the geodesics
of the spacetime to compute all non-vanishing Christoffel symbols for this metric.
(b) Consider the static limit of the line element (†) where λ and µ are functions of
the radius r only, and let the matter coupled to gravity be a spherically symmetric fluid
with energy momentum tensor

T µν = (ρ + P )uµ uν + P gµν , uµ = [λ−1 , 0, 0, 0] ,

where the pressure P and energy density ρ are also functions of the radius r. For
these Tolman-Oppenheimer-Volkoff stellar models, the Einstein and matter equations
Gµν = 8πTµν and ∇µ T µ ν = 0 reduce to

∂r λ µ2 − 1
= + 4πrµ2 P ,
λ 2r
 
r 1
∂r m = 4πr 2 ρ , where m(r) = 1− ,
2 µ2
 
µ2 − 1
∂r P = −(ρ + P ) + 4πrµ2 P .
2r

Consider now a constant density solution to the above Einstein and matter equa-
tions, where ρ takes the non-zero constant value ρ0 out to a radius R and ρ = 0 for r > R.
Show that for such a star,
 
4πr 1
∂r P = P + ρ0 (P + ρ0 ) ,
1 − 83 πρ0 r 2 3

and that the pressure at the centre of the star is

p
1 − 1 − 2M/R 4
P (0) = −ρ0 p , with M = πρ0 R3 .
3 1 − 2M/R − 1 3

Show that P (0) diverges
 if M = 4R/9. Hint: at the surface of the star the pressure
vanishes: P (R) = 0 .

Part II, Paper 4 [TURN OVER]

 26

37A Fluid Dynamics

(a) Show that the Stokes flow around a rigid moving sphere has the minimum viscous
dissipation rate of all incompressible flows which satisfy the no-slip boundary conditions
on the sphere.
(b) Let u = ∇(x · Φ + χ) − 2Φ, where Φ and χ are solutions of Laplace’s equation,
i.e. ∇2 Φ = 0 and ∇2 χ = 0.

(i) Show that u is incompressible.

(ii) Show that u satisfies Stokes equation if the pressure p = 2µ∇ · Φ.

(c) Consider a rigid sphere moving with velocity U . The Stokes flow around the
sphere is given by  
U 1
Φ=α and χ = βU · ∇ ,
r r
where the origin is chosen to be at the centre of the sphere. Find the values for α and β
which ensure no-slip conditions are satisfied on the sphere.

Part II, Paper 4

 27

38A Waves
(a) Assuming a slowly-varying two-dimensional wave pattern of the form
 
i
ϕ(x, t) = A(x, t; ε) exp θ(x, t) ,
ε

where 0 < ε ≪ 1, and a local dispersion relation ω = Ω(k; x, t), derive the ray tracing
equations,
dxi ∂Ω dω ∂Ω dki ∂Ω 1 dθ ∂Ω
= , = , =− , = −ω + kj ,
dt ∂ki dt ∂t dt ∂xi ε dt ∂kj

for i, j = 1, 2, explaining carefully the meaning of the notation used.

(b) For a homogeneous, time-independent (but not necessarily isotropic) medium,
show that all rays are straight lines. When the waves have zero frequency, deduce that if
the point x lies on a ray emanating from the origin in the direction given by a unit vector
b
cg , then
θ(x) = θ(0) + b
cg · k |x| .

(c) Consider a stationary obstacle in a steadily moving homogeneous medium which

has the dispersion relation

1/4
Ω = α k12 + k22 − V k1 ,
where (V, 0) is the velocity of the medium and α > 0 is a constant. The obstacle generates
a steady wave system. Writing (k1 , k2 ) = κ(cos φ, sin φ), with κ > 0, show that the wave
satisfies
α2
κ= 2 , bcg = (cos ψ, sin ψ) ,
V cos2 φ
where ψ is defined by
tan φ
tan ψ = −
1 + 2 tan2 φ
with 21 π < ψ < 32 π and − 21 π < φ < 21 π. Deduce that the wave pattern occupies a wedge


of semi-angle tan−1 2−3/2 , extending in the negative x1 -direction.

Part II, Paper 4 [TURN OVER]

 28

39C Numerical Analysis

For a 2-periodic analytic function f , its Fourier expansion is given by the formula
∞
X Z 1
1
f (x) = fbn eiπnx , fbn = f (t)e−iπnt dt .
n=−∞
2 −1

(a) Consider the two-point boundary value problem

X 2 ∞
1
− 2 (1 + 2 cos πx)u′′ + u = 1 + cos πnx, −1 6 x 6 1 ,
π n2 + 1
n=1

with periodic boundary conditions u(−1) = u(1). Construct explicitly the infinite
dimensional linear algebraic system that arises from the application of the Fourier spectral
method to the above equation, and explain how to truncate the system to a finite-
dimensional one.
(b) A rectangle rule is applied to computing the integral of a 2-periodic analytic
function h:
Z 1 N/2
X  
2 2k
h(t) dt ≈ h . (∗)
−1 N N
k=−N/2+1

Find an expression for the error eN (h) := LHS − RHS of (∗), in terms of b
hn , and show
that eN (h) has a spectral rate of decay as N → ∞. [In the last part, you may quote a
relevant theorem about bhn .]

END OF PAPER

Part II, Paper 4