MATHEMATICAL TRIPOS Part II

Tuesday, 4 June, 2019 1:30 pm to 4:30 pm MAT2

PAPER 2

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 a
Define the Jacobi symbol , where a, n ∈ Z and n is odd and positive.
n
State and prove the Law of Quadratic Reciprocity for the Jacobi symbol. [You
may use Quadratic Reciprocity for the Legendre symbol without proof but should state it
clearly.]
 
503
Compute the Jacobi symbol .
2019

2H Topics in Analysis
Let K be the collection of non-empty closed bounded subsets of Rn .
(a) Show that, if A, B ∈ K and we write

A + B = {a + b : a ∈ A, b ∈ B},

then A + B ∈ K.
(b) Show that, if Kn ∈ K, and

K1 ⊇ K2 ⊇ K3 ⊇ . . .
T∞
then K := n=1 Kn ∈ K.
(c) Assuming the result that

ρ(A, B) = sup inf |a − b| + sup inf |a − b|

a∈A b∈B b∈B a∈A

defines a metric on K (the Hausdorff metric), show that if Kn and K are as in part (b),
then ρ(Kn , K) → 0 as n → ∞.

3G Coding and Cryptography

Define the binary Hamming code of length n = 2l − 1 for l > 3. Define a perfect
code. Show that a binary Hamming code is perfect.
What is the weight of the dual code of a binary Hamming code when l = 3?

Part II, Paper 2

 3

4H Automata and Formal Languages

(a) Define a recursive set and a recursively enumerable (r.e.) set. Prove that E ⊆ N0
is recursive if and only if both E and N0 \ E are r.e. sets.
(b) Let E = {fn,k (m1 , . . . , mk ) | (m1 , . . . , mk ) ∈ Nk0 } for some fixed k > 1 and some
fixed register machine code n. Show that E = {m ∈ N0 | fj,1 (m) ↓} for some fixed register
machine code j. Hence show that E is an r.e. set.
(c) Show that the function f : N0 → N0 defined below is primitive recursive.

n − 1 if n > 0
f (n) =
0 if n = 0.

[Any use of Church’s thesis in your answers should be explicitly stated. In this question
N0 denotes the set of non-negative integers.]

Part II, Paper 2 [TURN OVER]

 4

5J Statistical Modelling
The cycling data frame contains the results of a study on the effects of cycling to
work among 1,000 participants with asthma, a respiratory illness. Half of the participants,
chosen uniformly at random, received a monetary incentive to cycle to work, and the other
half did not. The variables in the data frame are:

• miles: the average number of miles cycled per week

• episodes: the number of asthma episodes experienced during the study
• incentive: whether or not a monetary incentive to cycle was given
• history: the number of asthma episodes in the year preceding the study

Consider the R code below and its abbreviated output.

> lm.1 = lm(episodes ~ miles + history, data=cycling)

> summary(lm.1)

Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.66937 0.07965 8.404 < 2e-16 ***
miles -0.04917 0.01839 -2.674 0.00761 **
history 1.48954 0.04818 30.918 < 2e-16 ***

> lm.2 = lm(episodes ~ incentive + history, data=cycling)

> summary(lm.2)

Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.09539 0.06960 1.371 0.171
incentiveYes 0.91387 0.06504 14.051 <2e-16 ***
history 1.46806 0.04346 33.782 <2e-16 ***

> lm.3 = lm(miles ~ incentive + history, data=cycling)

> summary(lm.3)

Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 1.47050 0.11682 12.588 < 2e-16 ***
incentiveYes 1.73282 0.10917 15.872 < 2e-16 ***
history 0.47322 0.07294 6.487 1.37e-10 ***

(a) For each of the fitted models, briefly explain what can be inferred about
participants with similar histories.
(b) Based on this analysis and the experimental design, is it advisable for a
participant with asthma to cycle to work more often? Explain.

Part II, Paper 2

 5

6C Mathematical Biology
An activator–inhibitor system for u(x, t) and v(x, t) is described by the equations

∂u ∂2u
= uv 2 − a + D ,
∂t ∂x2
∂v ∂2v
= v − uv 2 + 2 ,
∂t ∂x
where a, D > 0.
Find the range of a for which the spatially homogeneous system has a stable
equilibrium solution with u > 0 and v > 0.
For the case when the homogeneous system is stable, consider spatial perturbations
proportional to cos(kx) to the equilibrium solution found above. Give a condition on D
in terms of a for the system to have a Turing instability (a spatial instability).

7A Further Complex Methods

Assume that |f (z)/z| → 0 as |z| → ∞ and that f (z) is analytic in the upper
half-plane (including the real axis). Evaluate
Z ∞
f (x)
P 2 2
dx,
−∞ x(x + a )

where a is a positive real number.

[You must state clearly any standard results involving contour integrals that you use.]

Part II, Paper 2 [TURN OVER]

 6

8E Classical Dynamics
(a) State Hamilton’s equations for a system with n degrees of freedom and Hamilto-
nian H(q, p, t), where (q, p) = (q1 , . . . , qn , p1 , . . . , pn ) are canonical phase-space variables.
(b) Define the Poisson bracket {f, g} of two functions f (q, p, t) and g(q, p, t).
(c) State the canonical commutation relations of the variables q and p.
(d) Show that the time-evolution of any function f (q, p, t) is given by

df ∂f
= {f, H} + .
dt ∂t

(e) Show further that the Poisson bracket of any two conserved quantities is also a
conserved quantity.
[You may assume the Jacobi identity,

{f, {g, h}} + {g, {h, f }} + {h, {f, g}} = 0 . ]

9B Cosmology
[You may work in units of the speed of light, so that c = 1.]
(a) Combining the Friedmann and continuity equations

8πG
H2 = ρ, ρ̇ + 3H(ρ + P ) = 0 ,
3
derive the Raychaudhuri equation (also known as the acceleration equation) which expresses
ä/a in terms of the energy density ρ and the pressure P .
(b) Assuming an equation of state P = wρ with constant w, for what w is the
expansion of the universe accelerated or decelerated?
(c) Consider an expanding, spatially-flat FLRW universe with both a cosmological
constant and non-relativistic matter (also known as dust) with energy densities ρcc and
ρdust respectively. At some time corresponding to aeq , the energy densities of these two
components are equal ρcc (aeq ) = ρdust (aeq ). Is the expansion of the universe accelerated
or decelerated at this time?
(d) For what numerical value of a/aeq does the universe transition from deceleration
to acceleration?

Part II, Paper 2

 7

10D Quantum Information and Computation

The BB84 quantum key distribution protocol begins with Alice choosing two
uniformly random bit strings X = x1 x2 . . . xm and Y = y1 y2 . . . ym .
(a) In terms of these strings, describe Alice’s process of conjugate coding for the
BB84 protocol.
(b) Suppose Alice and Bob are distantly separated in space and have available a
noiseless quantum channel on which there is no eavesdropping. They can also communicate
classically publicly. For this idealised situation, describe the steps of the BB84 protocol
that results in Alice and Bob sharing a secret key of expected length m/2.
(c) Suppose now that an eavesdropper Eve taps into the channel and carries out the
following action on each passing qubit. With probability 1−p, Eve lets it pass undisturbed,
and with probability p she chooses a bit w ∈ {0, 1} uniformly at random√ and measures
√
the qubit in basis Bw where B0 = {|0i , |1i} and B1 = {(|0i + |1i)/ 2, (|0i − |1i)/ 2}.
After measurement Eve sends the post-measurement state on to Bob. Calculate the bit
error rate for Alice and Bob’s final key in part (b) that results from Eve’s action.

Part II, Paper 2 [TURN OVER]

 8

SECTION II
11H Topics in Analysis
Throughout this question I denotes the closed interval [−1, 1].
(a) For n ∈ N, consider the 2n + 1 points r/n ∈ I with r ∈ Z and −n 6 r 6 n. Show
that, if we colour them red or green in such a way that −1 and 1 are coloured differently,
there must be two neighbouring points of different colours.
(b) Deduce from part (a) that, if I = A ∪ B with A and B closed, −1 ∈ A and
1 ∈ B, then A ∩ B 6= ∅.
(c) Deduce from part (b) that there does not exist a continuous function f : I → R
with f (t) ∈ {−1, 1} for all t ∈ I and f (−1) = −1, f (1) = 1.
(d) Deduce from part (c) that if f : I → I is continuous then there exists an x ∈ I
with f (x) = x.
(e) Deduce the conclusion of part (c) from the conclusion of part (d).
(f) Deduce the conclusion of part (b) from the conclusion of part (c).
(g) Suppose that we replace I wherever it occurs by the unit circle

C = {z ∈ C | |z| = 1}.

Which of the conclusions of parts (b), (c) and (d) remain true? Give reasons.

12G Coding and Cryptography

Describe the Huffman coding scheme and prove that Huffman codes are optimal.
Are the following statements true or false? Justify your answers.
(i) Given m messages with probabilities p1 > p2 > · · · > pm a Huffman coding will
assign a unique set of word lengths.
(ii) An optimal code must be Huffman.
(iii) Suppose the m words of a Huffman code have word lengths s1 , s2 , . . . , sm . Then
m
X
2−si = 1.
i=1

[Throughout this question you may assume that a decipherable code with prescribed
word lengths exists if and only if there is a prefix-free code with the same word lengths.]

Part II, Paper 2

 9

13A Further Complex Methods

The Riemann zeta function is defined as
∞
X 1
ζ(z) := (†)
nz
n=1

for Re(z) > 1, and by analytic continuation to the rest of C except at singular points.
The integral representation of (†) for Re(z) > 1 is given by
Z ∞ z−1
1 t
ζ(z) = dt (‡)
Γ(z) 0 et − 1

where Γ is the Gamma function.

(a) The Hankel representation is defined as
Z (0+ )
Γ(1 − z) tz−1
ζ(z) = dt . (⋆)
2πi −∞ e−t − 1

Explain briefly why this representation gives an analytic continuation of ζ(z) as defined
in (‡) to all z other than z = 1, using a diagram to illustrate what is meant by the upper
limit of the integral in (⋆).
[You may assume Γ(z)Γ(1 − z) = π/ sin(πz).]
(b) Find
 
t−z
Res , t = 2πin ,
e−t − 1
where n is an integer and the poles are simple.
(c) By considering Z
t−z
dt ,
γ e−t − 1
where γ is a suitably modified Hankel contour and using the result of part (b), derive the
reflection formula:

ζ(1 − z) = 21−z π −z cos 1
2 πz Γ(z)ζ(z) .

Part II, Paper 2 [TURN OVER]

 10

14E Classical Dynamics

The Lagrangian of a particle of mass m and charge q moving in an electromagnetic
field described by scalar and vector potentials φ(r, t) and A(r, t) is
1
L= m|ṙ|2 + q(−φ + ṙ · A) ,
2
where r(t) is the position vector of the particle and ṙ = dr/dt.
(a) Show that Lagrange’s equations are equivalent to the equation of motion

mr̈ = q(E + v × B) ,

where
∂A
E = −∇φ − , B=∇×A
∂t
are the electric and magnetic fields.
(b) Show that the related Hamiltonian is

|p − qA|2
H= + qφ ,
2m
where p = mṙ + qA. Obtain Hamilton’s equations for this system.
(c) Verify that the electric and magnetic fields remain unchanged if the scalar and
vector potentials are transformed according to
∂f
φ 7→ φ̃ = φ −,
∂t
A 7→ Ã = A + ∇f ,

where f (r, t) is a scalar field. Show that the transformed Lagrangian L̃ differs from L by
the total time-derivative of a certain quantity. Why does this leave the form of Lagrange’s
equations invariant? Show that the transformed Hamiltonian H̃ and phase-space variables
(r, p̃) are related to H and (r, p) by a canonical transformation.
[Hint: In standard notation, the canonical transformation associated with the type-2
generating function F2 (q, P, t) is given by

∂F2 ∂F2 ∂F2

p= , Q= , K=H+ .]
∂q ∂P ∂t

Part II, Paper 2

 11

15D Quantum Information and Computation

Let |α0 i =
6 |α1 i be two quantum states and let p0 and p1 be associated probabilities
with p0 + p1 = 1, p0 6= 0, p1 6= 0 and p0 > p1 . Alice chooses state |αi i with probability pi
and sends it to Bob. Upon receiving it, Bob performs a 2-outcome measurement M with
outcomes labelled 0 and 1, in an attempt to identify which state Alice sent.
(a) By using the extremal property of eigenvalues, or otherwise, show that the
operator D = p0 |α0 i hα0 | − p1 |α1 i hα1 | has exactly two nonzero eigenvalues, one of which
is positive and the other negative.
(b) Let PS denote the probability that Bob correctly identifies Alice’s sent state. If
the measurement M comprises orthogonal projectors {Π0 , Π1 } (corresponding to outcomes
0 and 1 respectively) give an expression for PS in terms of p1 , Π0 and D.
(c) Show that the optimal success probability PSopt , i.e. the maximum attainable
value of PS , is p
opt 1 + 1 − 4p0 p1 cos2 θ
PS = ,
2
where cos θ = |hα0 |α1 i|.
(d) Suppose we now place the following extra requirement on Bob’s discrimination
process: whenever Bob obtains output 0 then the state sent by Alice was definitely |α0 i.
Show that Bob’s PSopt now satisfies PSopt > 1 − p0 cos2 θ.

16I Logic and Set Theory

Give the inductive and synthetic definitions of ordinal addition, and prove that they
are equivalent.
Which of the following assertions about ordinals α, β and γ are always true, and
which can be false? Give proofs or counterexamples as appropriate.
(i) α + (β + γ) = (α + β) + γ.
(ii) If α and β are uncountable then α + β = β + α.
(iii) α(βγ) = (αβ)γ.
(iv) If α and β are infinite and α + β = β + α then αβ = βα.

Part II, Paper 2 [TURN OVER]

 12

17G Graph Theory

(a) Suppose that the edges of the complete graph K6 are coloured blue and yellow.
Show that it must contain a monochromatic triangle. Does this remain true if K6 is
replaced by K5 ?
(b) Let t > 1. Suppose that the edges of the complete graph K3t−1 are coloured
blue and yellow. Show that it must contain t edges of the same colour with no two sharing
a vertex. Is there any t > 1 for which this remains true if K3t−1 is replaced by K3t−2 ?
(c) Now let t > 2. Suppose that the edges of the complete graph Kn are coloured
blue and yellow in such a way that there are a blue triangle and a yellow triangle with
no vertices in common. Show that there are also a blue triangle and a yellow triangle
that do have a vertex in common. Hence, or otherwise, show that whenever the edges of
the complete graph K5t are coloured blue and yellow it must contain t monochromatic
triangles, all of the same colour, with no two sharing a vertex. Is there any t > 2 for
which this remains true if K5t is replaced by K5t−1 ? [You may assume that whenever
the edges of the complete graph K10 are coloured blue and yellow it must contain two
monochromatic triangles of the same colour with no vertices in common.]

18F Galois Theory

For any prime p 6= 5, explain briefly why the Galois group of X 5 − 1 over Fp is cyclic
of order d, where d = 1 if p ≡ 1 mod 5, d = 4 if p ≡ 2, 3 mod 5, and d = 2 if p ≡ 4
mod 5.
Show that the splitting field of X 5 − 5 over Q is an extension of degree 20.
For any prime p 6= 5, prove that X 5 − 5 ∈ Fp [X] does not have an irreducible cubic
as a factor. For p ≡ 2 or 3 mod 5, show that X 5 − 5 is the product of a linear factor and
an irreducible quartic over Fp . For p ≡ 1 mod 5, show that either X 5 − 5 is irreducible
over Fp or it splits completely.
[You may assume the reduction mod p criterion for finding cycle types in the Galois
group of a monic polynomial over Z and standard facts about finite fields.]

Part II, Paper 2

 13

19I Representation Theory

(a) For any finite group G, let ρ1 , . . . , ρk be a complete set of non-isomorphic complex
irreducible representations of G, with dimensions n1 , . . . nk , respectively. Show that
k
X
n2j = |G|.
j=1

(b) Let A, B, C, D be the matrices

   
−1 0 0 0 0 0 i 0
 0 1 0 0   0 0 0 −i 
A=
 0
, B =  ,
0 1 0   i 0 0 0 
0 0 0 −1 0 −i 0 0
   
0 i 0 0 0 0 0 1
 i 0 0 
0   0 0 1 0 
C=
 0 , D= ,
0 0 −i   0 1 0 0 
0 0 −i 0 1 0 0 0
and let G = hA, B, C, Di. Write Z = −I4 .
(i) Prove that the derived subgroup G′ = hZi.
(ii) Show that for all g ∈ G, g2 ∈ hZi, and deduce that G is a 2-group of order at
most 32.
(iii) Prove that the given representation of G of degree 4 is irreducible.
(iv) Prove that G has order 32, and find all the irreducible representations of G.

20G Number Fields

(a) Let L be a number field. State Minkowski’s upper bound for the norm of a
representative for a given class of the ideal class group Cl(OL ).
√ √
(b) Now let K = Q( −47) and ω = 21 (1 + −47). Using Dedekind’s criterion, or
otherwise, factorise the ideals (ω) and (2 + ω) as products of non-zero prime ideals of OK .
(c) Show that Cl(OK ) is cyclic, and determine its order.
[You may assume that OK = Z[ω].]

Part II, Paper 2 [TURN OVER]

 14

21F Algebraic Topology

Let T = S 1 × S 1 , U = S 1 × D 2 and V = D 2 × S 1 . Let i : T → U , j : T → V be the
natural inclusion maps. Consider the space S := U ∪T V ; that is,

S := (U ⊔ V )/ ∼

where ∼ is the smallest equivalence relation such that i(x) ∼ j(x) for all x ∈ T .
(a) Prove that S is homeomorphic to the 3-sphere S 3 .
[Hint: It may help to think of S 3 as contained in C2 .]
(b) Identify T as a quotient of the square I × I in the usual way. Let K be the circle
in T given by the equation y = 32 x mod 1. K is illustrated in the figure below.

Compute a presentation for π1 (S − K), where S − K is the complement of K in S,

and deduce that π1 (S − K) is non-abelian.

Part II, Paper 2

 15

22H Linear Analysis

(a) State the real version of the Stone–Weierstrass theorem and state the Urysohn–
Tietze extension theorem.
(b) In this part, you may assume that there is a sequence of polynomials Pi such
√
that supx∈[0,1] |Pi (x) − x| → 0 as i → ∞.
Let f : [0, 1] → R be a continuous piecewise linear function which is linear on
[0, 1/2] and on [1/2, 1]. Using the polynomials Pi mentioned above (but not assuming any
form of the Stone-Weierstrass theorem), prove that there are polynomials Qi such that
supx∈[0,1] |Qi (x) − f (x)| → 0 as i → ∞.
(d) Which of the following families of functions are relatively compact in C[0, 1]
with the supremum norm? Justify your answer.

sin(πnx)
F1 = {x 7→ : n ∈ N}
n
sin(πnx)
F2 = {x 7→ : n ∈ N}
n1/2
F3 = {x 7→ sin(πnx) : n ∈ N}

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

23F Riemann Surfaces

(a) Prove that z 7→ z 4 as a map from the upper half-plane H to C \ {0} is a covering
map which is not regular.
(b) Determine the set of singular points on the unit circle for
∞
X
h(z) = (−1)n (2n + 1)z n .
n=0

(c) Suppose f : ∆ \ {0} → ∆ \ {0} is a holomorphic map where ∆ is the unit disk.
Prove that f extends to a holomorphic map f˜ : ∆ → ∆. If additionally f is biholomorphic,
prove that f˜(0) = 0.
(d) Suppose that g : C ֒→ R is a holomorphic injection with R a compact Riemann
surface. Prove that R has genus 0, stating carefully any theorems you use.

Part II, Paper 2 [TURN OVER]

 16

24F Algebraic Geometry

(a) Let A be a commutative algebra over a field k, and p : A → k a k-linear
homomorphism. Define Der(A, p), the derivations of A centered in p, and define the
tangent space Tp A in terms of this.
Show directly from your definition that if f ∈ A is not a zero divisor and p(f ) 6= 0,
then the natural map Tp A[ f1 ] → Tp A is an isomorphism.
(b) Suppose k is an algebraically closed field and λi ∈ k for 1 6 i 6 r. Let

X = {(x, y) ∈ A2 | x 6= 0, y 6= 0, y 2 = (x − λ1 ) · · · (x − λr )}.

Find a surjective map X → A1 . Justify your answer.

25H Differential Geometry

(a) Let α : (a, b) → R3 be a smooth regular curve parametrised by arclength. For
s ∈ (a, b), define the curvature k(s) and (where defined) the torsion τ (s) of α. What
condition must be satisfied in order for the torsion to be defined? Derive the Frenet
equations.
(b) If τ (s) is defined and equal to 0 for all s ∈ (a, b), show that α lies in a plane.
(c) State the fundamental theorem for regular curves in R3 , giving necessary and
sufficient conditions for when curves α(s) and α e(s) are related by a proper Euclidean
motion.
(d) Now suppose that αe : (a, b) → R3 is another smooth regular curve parametrised
by arclength, and that ek(s) and τe(s) are its curvature and torsion. Determine whether
the following statements are true or false. Justify your answer in each case.

(i) If τ (s) = 0 whenever it is defined, then α lies in a plane.

(ii) If τ (s) is defined and equal to 0 for all but one value of s in (a, b), then α lies
in a plane.

(iii) If k(s) = e k(s) for all s, τ (s) and τe(s) are defined for all s 6= s0 , and τ (s) = τe(s)
for all s 6= s0 , then α and α e are related by a rigid motion.

Part II, Paper 2

 17

26K Probability and Measure

(a) Let (Xi , Ai ) for i = 1, 2 be two measurable spaces. Define the product σ-algebra
A1 ⊗ A2 on the Cartesian product X1 × X2 . Given a probability measure µi on (Xi , Ai )
for each i = 1, 2, define the product measure µ1 ⊗ µ2 . Assuming the existence of a product
measure, explain why it is unique. [You may use standard results from the course if clearly
stated.]
(b) Let (Ω, F, P) be a probability space on which the real random variables U and V
are defined. Explain what is meant when one says that U has law µ. On what measurable
space is the measure µ defined? Explain what it means for U and V to be independent
random variables.
(c) Now let X = [− 21 , 21 ], let A be its Borel σ-algebra and let µ be Lebesgue
measure. Give an example of a measure η on the product (X × X, A ⊗ A) such that
η(X × A) = µ(A) = η(A × X) for every Borel set A, but such that η is not Lebesgue
measure on X × X.
(d) Let η be as in part (c) and let I, J ⊂ X be intervals of length x and y respectively.
Show that
x + y − 1 6 η(I × J) 6 min{x, y}.

(e) Let X be as in part (c). Fix d > 2 and let Πi denote the projection
Πi (x1 , . . . , xd ) = (x1 , . . . , xi−1 , xi+1 , . . . , xd ) from X d to X d−1 . Construct a probability
measure η on X d , such that the image under each Πi coincides with the (d−1)-dimensional
Lebesgue measure, while η itself is not the d-dimensional Lebesgue measure. [Hint:
Consider the following collection of 2d − 1 independent random variables: U1 , . . . , Ud
uniformly distributed on [0, 12 ], and ε1 , . . . , εd−1 such that P(εi = 1) = P(εi = −1) = 21 for
each i.]

Part II, Paper 2 [TURN OVER]

 18

27K Applied Probability

Let X = (Xt : t > 0) be a Markov chain on the non-negative integers with generator
G = (gi,j ) given by

gi,i+1 = λi , i > 0,
gi,0 = λi ρi , i > 0,
gi,j = 0, j 6= 0, i, i + 1,

for a given collection of positive numbers λi , ρi .

(a) State the transition matrix of the jump chain Y of X.
(b) Why is X not reversible?
Q
(c) Prove that X is transient if and only if i (1 + ρi ) < ∞.
Q
(d) Assume that i (1 + ρi ) < ∞. Derive a necessary and sufficient condition on
the parameters λi , ρi for X to be explosive.
(e) Derive a necessary and sufficient condition on the parameters λi , ρi for the
existence of an invariant measure for X.
[You may use any general results from the course concerning Markov chains and
pure birth processes so long as they are clearly stated.]

Part II, Paper 2

 19

28J Principles of Statistics

(a) We consider the model {P oisson(θ) : θ ∈ (0, ∞)} and an i.i.d. sample
X1 , . . . , Xn from it. Compute the expectation and variance of X1 and check they are
equal. Find the maximum likelihood estimator θ̂M LE for θ and, using its form, derive the
√
limit in distribution of n(θ̂M LE − θ).
(b) In practice, Poisson-looking data show overdispersion, i.e., the sample variance
is larger than the sample expectation. For π ∈ [0, 1] and λ ∈ (0, ∞), let pπ,λ : N0 → [0, 1],
 k
 πe−λ λk! for k > 1
k 7→ pπ,λ (k) =

(1 − π) + πe−λ for k = 0.

Show that this defines a distribution. Does it model overdispersion? Justify your answer.
(c) Let Y1 , . . . , Yn be an i.i.d. sample from pπ,λ . Assume λ is known. Find the
maximum likelihood estimator π̂M LE for π.
√
(d) Furthermore, assume that, for any π ∈ [0, 1], n(π̂M LE − π) converges in
distribution to a random variable Z as n → ∞. Suppose we wanted to test the null
hypothesis that our data arises from the model in part (a). Before making any further
computations, can we necessarily expect Z to follow a normal distribution under the null
hypothesis? Explain. Check your answer by computing the appropriate distribution.
[You may use results from the course, provided you state it clearly.]

Part II, Paper 2 [TURN OVER]

 20

29K Stochastic Financial Models

(a) In the context of a multi-period model in discrete time, what does it mean to
say that a probability measure is an equivalent martingale measure?
(b) State the fundamental theorem of asset pricing.
(c) Consider a single-period model with one risky asset S 1 having initial price S01 = 1.
At time 1 its value S11 is a random variable on (Ω, F, P) of the form


S11 = exp σZ + m , m ∈ R, σ > 0,

where Z ∼ N (0, 1). Assume that there is a riskless numéraire S 0 with S00 = S10 = 1. Show
that there is no arbitrage in this model.
[Hint: You may find it useful to consider a density of the form exp(σ̃Z + m̃) and
find suitable m̃ and σ̃. You may use
without proof that if X is a normal random variable
then E(eX ) = exp E(X) + 21 Var(X) .]
(d) Now consider a multi-period model with one risky asset S 1 having a non-random
initial price S01 = 1 and a price process (St1 )t∈{0,...,T } of the form

t
Y

St1 = exp σi Zi + mi , mi ∈ R, σi > 0,
i=1

where Zi are i.i.d. N (0, 1)-distributed random variables on (Ω, F, P). Assume that there
is a constant riskless numéraire S 0 with St0 = 1 for all t ∈ {0, . . . , T }. Show that there
exists no arbitrage in this model.

Part II, Paper 2

 21

30A Asymptotic Methods

(a) Define formally what it means for a real valued function f (x) to have an
asymptotic expansion about x0 , given by
∞
X
f (x) ∼ fn (x − x0 )n as x → x0 .
n=0

Use this definition to prove the following properties.

(i) If both f (x) and g(x) have asymptotic expansions about x0 , then
h(x) = f (x) + g(x) also has an asymptotic expansion about x0 .

(ii) If f (x) has an asymptotic expansion about x0 and is integrable, then

Z x ∞
X fn
f (ξ) dξ ∼ (x − x0 )n+1 as x → x0 .
x0 n+1
n=0

(b) Obtain, with justification, the first three terms in the asymptotic expansion as
x → ∞ of the complementary error function, erfc(x), defined as
Z ∞
1 2
erfc(x) := √ e−t dt.
2π x

31E Dynamical Systems

For a map F : Λ → Λ give the definitions of chaos according to (i) Devaney (D-
chaos) and (ii) Glendinning (G-chaos).
Consider the dynamical system

F (x) = ax (mod 1)

on Λ = [0, 1), for a > 1 (note that a is not necessarily an integer). For both definitions of
chaos, show that this system is chaotic.

Part II, Paper 2 [TURN OVER]

 22

32C Integrable Systems

Suppose p = p(x) is a smooth, real-valued, function of x ∈ R which satisfies p(x) > 0
for all x and p(x) → 1, px (x), pxx (x) → 0 as |x| → ∞. Consider the Sturm-Liouville
operator:  
d 2 dψ
Lψ := − p ,
dx dx
which acts on smooth, complex-valued, functions ψ = ψ(x). You may assume that for any
k > 0 there exists a unique function ϕk (x) which satisfies:

Lϕk = k2 ϕk ,

and has the asymptotic behaviour:

 −ikx
e as x → −∞,
ϕk (x) ∼
a(k)e−ikx + b(k)eikx as x → +∞.

(a) By analogy with the standard Schrödinger scattering problem, define the
reflection and transmission coefficients: R(k), T (k). Show that |R(k)|2 + |T (k)|2 = 1.
[Hint: You may wish to consider W (x) = p(x)2 [ψ1 (x)ψ2′ (x) − ψ2 (x)ψ1′ (x)] for suitable
functions ψ1 and ψ2 .]
(b) Show that, if κ > 0, there exists no non-trivial normalizable solution ψ to the
equation
Lψ = −κ2 ψ.

Assume now that p = p(x, t), such that p(x, t) > 0 and p(x, t) → 1, px (x, t), pxx (x, t) →
0 as |x| → ∞. You are given that the operator A defined by:

d3 ψ d2 ψ dψ
Aψ := −4p3 3
− 18p 2
p x 2
− (12pp2x + 6p2 pxx ) ,
dx dx dx
satisfies:  
d 4 dψ
(LA − AL)ψ = − 2p pxxx .
dx dx

(c) Show that L, A form a Lax pair if the Harry Dym equation,

pt = p3 pxxx

is satisfied. [You may assume L = L† , A = −A† .]

(d) Assuming that p solves the Harry Dym equation, find how the transmission and
reflection amplitudes evolve as functions of t.

Part II, Paper 2

 23

33B Principles of Quantum Mechanics

(a) Let |ii and |ji be two eigenstates of a time-independent Hamiltonian H0 ,
separated in energy by ~ωij . At time t = 0 the system is perturbed by a small, time
independent operator V . The perturbation is turned off at time t = T . Show that if the
system is initially in state |ii, the probability of a transition to state |ji is approximately

sin2 (ωij T /2)

Pij = 4|hi|V |ji|2 .
(~ωij )2

(b) An uncharged particle with spin one-half and magnetic moment µ travels at
speed v through a region of uniform magnetic field B = (0, 0, B). Over a length L of its
path, an additional perpendicular magnetic field b is applied. The spin-dependent part of
the Hamiltonian is
(
−µ(Bσz + bσx ) while 0 < t < L/v
H(t) =
−µBσz otherwise,

where σz and σx are Pauli matrices. The particle initially has its spin aligned along the
direction of B = (0, 0, B). Find the probability that it makes a transition to the state
with opposite spin

(i) by assuming b ≪ B and using your result from part (a),

(ii) by finding the exact evolution of the state.

[Hint: for any 3-vector a, eia·σ = (cos a)I + (i sin a) â · σ, where I is the 2 × 2 unit matrix,
σ = (σx , σy , σz ), a=|a| and â = a/|a|. ]

34B Applications of Quantum Mechanics

Give an account of the variational principle for establishing an upper bound on the
ground state energy of a Hamiltonian H.
A particle of mass m moves in one dimension and experiences the potential
V = A|x|n with n an integer. Use a variational argument to prove the virial theorem,

2hT i0 = nhV i0

where h·i0 denotes the expectation value in the true ground state. Deduce that there is
no normalisable ground state for n 6 −3.
2 x2
For the case n = 1, use the ansatz ψ(x) ∝ e−α to find an estimate for the energy
of the ground state.

Part II, Paper 2 [TURN OVER]

 24

35D Statistical Physics

Using the classical statistical mechanics of a gas of molecules with negligible
interactions, derive the ideal gas law. Explain briefly to what extent this law is independent
of the molecule’s internal structure.
Calculate the entropy S of a monatomic gas of low density, with negligible interac-
tions. Deduce the equation relating the pressure P and volume V of the gas on a curve
in the P V -plane along which S is constant.
Z ∞ π 1
2
e−αx dx =
2
[You may use for α > 0 .]
−∞ α

36D General Relativity

Consider the spacetime metric
1 2m
ds2 = −f (r)dt2 + dr 2 + r 2 (dθ 2 + sin2 θ dφ2 ) , with f (r) = 1 − − H 2 r2 ,
f (r) r

where H > 0 and m > 0 are constants.

(a) Write down the Lagrangian for geodesics in this spacetime, determine three
independent constants of motion and show that geodesics obey the equation

ṙ 2 + V (r) = E 2 ,

where E is constant, the overdot denotes differentiation with respect to an affine parameter
and V (r) is a potential function to be determined.
(b) Sketch the potential V (r) for the case of null geodesics, find any circular null
geodesics of this spacetime, and determine whether they are stable or unstable.
√
(c) Show that f (r) has√two positive roots r− and r+ if mH < 1/ 27 and that these
satisfy the relation r− < 1/( 3H) < r+ .
(d) Describe in one sentence the physical significance of those points where f (r) = 0.

Part II, Paper 2

 25

37A Fluid Dynamics

A viscous fluid is contained in a channel between rigid planes y = −h and y = h.
The fluid in the upper region σ < y < h (with −h < σ < h) has dynamic viscosity µ−
while the fluid in the lower region −h < y < σ has dynamic viscosity µ+ > µ− . The plane
at y = h moves with velocity U− and the plane at y = −h moves with velocity U+ , both
in the x direction. You may ignore the effect of gravity.

(a) Find the steady, unidirectional solution of the Navier-Stokes equations in which
the interface between the two fluids remains at y = σ.

(b) Using the solution from part (a):

(i) calculate the stress exerted by the fluids on the two boundaries;

(ii) calculate the total viscous dissipation rate in the fluids;

(iii) demonstrate that the rate of working by boundaries balances the viscous
dissipation rate in the fluids.

(c) Consider the situation where U+ + U− = 0. Defining the volume flux in the
upper region as Q− and the volume flux in the lower region as Q+ , show that their ratio
is independent of σ and satisfies
Q− µ−
=− .
Q+ µ+

Part II, Paper 2 [TURN OVER]

 26

38A Waves
The linearised equation of motion governing small disturbances in a homogeneous
elastic medium of density ρ is

∂2u
ρ = (λ + µ)∇(∇ · u) + µ∇2 u ,
∂t2
where u(x, t) is the displacement, and λ and µ are the Lamé moduli.
(a) The medium occupies the region between a rigid plane boundary at y = 0 and
a free surface at y = h. Show that SH waves can propagate in the x-direction within this
region, and find the dispersion relation for such waves.
(b) For each mode, deduce the cutoff frequency, the phase velocity and the group
velocity. Plot the latter two velocities as a function of wavenumber.
(c) Verify that in an average sense (to be made precise), the wave energy flux is
equal to the wave energy density multiplied by the group velocity.
[You may assume that the elastic energy per unit volume is given by

Ep = 12 λeii ejj + µeij eij .]

Part II, Paper 2

 27

39C Numerical Analysis

The Poisson equation on the unit square, equipped with zero boundary conditions,
is discretized with the 9-point scheme:

− 10 2
3 ui,j + 3 (ui+1,j + ui−1,j + ui,j+1 + ui,j−1 )
+ 16 (ui+1,j+1 + ui+1,j−1 + ui−1,j+1 + ui−1,j−1 ) = h2 fi,j ,
1
where 1 6 i, j 6 m, ui,j ≈ u(ih, jh), and (ih, jh) are the grid points with h = m+1 . We
also assume that u0,j = ui,0 = um+1,j = ui,m+1 = 0.
(a) Prove that all m × m tridiagonal symmetric Toeplitz (TST-) matrices
α β
 
..
 β α .
 ∈ Rm×m

H = [β, α, β] :=  .. .. (1)
 . . β 
β α

share the same eigenvectors q k with the components (sin jkπh)m j=1 for k = 1, ..., m.
Find expressions for the corresponding eigenvalues λk for k = 1, ..., m. Deduce that
H = QDQ−1 , where D = diag{λk } and Q is the matrix (sin ijπh)m
i,j=1 .

(b) Show that, by arranging the grid points (ih, jh) into a one-dimensional array
by columns, the 9-points scheme results in the following system of linear equations of the
form   u1   b1 
B C

.
 C B ..  u2   b2 
Au = b .. ..  . = . , (2)
  
⇔ 
 . . C  ..
   .. 
C B um bm
2 2 2
where A ∈ Rm ×m , the vectors uk , bk ∈ Rm are portions of u, b ∈ Rm , respectively, and
B, C are m × m TST-matrices whose elements you should determine.
(c) Using v k = Q−1 uk , ck = Q−1 bk , show that (2) is equivalent to
  v1  
c1

D E

.
 E D ..  v2   c2 
. .  . = .. , (3)
  
. .

. . E   ..

.
   
E D vm cm

where D and E are diagonal matrices.

(d) Show that, by appropriate reordering of the grid, the system (3) is reduced to
m uncoupled m × m systems of the form

Λk vbk = b
ck , k = 1, . . . , m.

Determine the elements of the matrices Λk .

END OF PAPER

Part II, Paper 2