Calculus
Alan Bain
1. Introduction
The following notes aim to provide a very informal introduction to Stochastic Calculus,
and especially to the Ito integral and some of its applications. They owe a great deal to Dan
Crisans Stochastic Calculus and Applications lectures of 1998; and also much to various
books especially those of L. C. G. Rogers and D. Williams, and Dellacherie and Meyers
multi volume series Probabilities et Potentiel. They have also beneted from insights
gained by attending lectures given by T. Kurtz.
The present notes grew out of a set of typed notes which I produced when revising
for the Cambridge, Part III course; combining the printed notes and my own handwritten
notes into a consistent text. Ive subsequently expanded them inserting some extra proofs
from a great variety of sources. The notes principally concentrate on the parts of the course
which I found hard; thus there is often little or no comment on more standard matters; as
a secondary goal they aim to present the results in a form which can be readily extended
Due to their evolution, they have taken a very informal style; in some ways I hope this
may make them easier to read.
The addition of coverage of discontinuous processes was motivated by my interest in
the subject, and much insight gained from reading the excellent book of J. Jacod and
A. N. Shiryaev.
The goal of the notes in their current form is to present a fairly clear approach to
the Ito integral with respect to continuous semimartingales but without any attempt at
maximal detail. The various alternative approaches to this subject which can be found
in books tend to divide into those presenting the integral directed entirely at Brownian
Motion, and those who wish to prove results in complete generality for a semimartingale.
Here at all points clarity has hopefully been the main goal here, rather than completeness;
although secretly the approach aims to be readily extended to the discontinuous theory.
I make no apology for proofs which spell out every minute detail, since on a rst look at
the subject the purpose of some of the steps in a proof often seems elusive. Id especially
like to convince the reader that the Ito integral isnt that much harder in concept than
the Lebesgue Integral with which we are all familiar. The motivating principle is to try
and explain every detail, no matter how trivial it may seem once the subject has been
understood!
Passages enclosed in boxes are intended to be viewed as digressions from the main
text; usually describing an alternative approach, or giving an informal description of what
is going on feel free to skip these sections if you nd them unhelpful.
In revising these notes I have resisted the temptation to alter the original structure
of the development of the Ito integral (although I have corrected unintentional mistakes),
since I suspect the more concise proofs which I would favour today would not be helpful
on a rst approach to the subject.
These notes contain errors with probability one. I always welcome people telling me
about the errors because then I can x them! I can be readily contacted by email as
alanb@chiark.greenend.org.uk. Also suggestions for improvements or other additions
are welcome.
Alan Bain
[i]
2. Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . i
2. Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ii
3. Stochastic Processes . . . . . . . . . . . . . . . . . . . . . . . 1
3.1. Probability Space . . . . . . . . . . . . . . . . . . . . . . . 1
3.2. Stochastic Process . . . . . . . . . . . . . . . . . . . . . . 1
4. Martingales . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
4.1. Stopping Times . . . . . . . . . . . . . . . . . . . . . . . 4
5. Basics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
5.1. Local Martingales . . . . . . . . . . . . . . . . . . . . . . 8
5.2. Local Martingales which are not Martingales . . . . . . . . . . . 9
6. Total Variation and the Stieltjes Integral . . . . . . . . . . . . . 11
6.1. Why we need a Stochastic Integral . . . . . . . . . . . . . . . 11
6.2. Previsibility . . . . . . . . . . . . . . . . . . . . . . . . . 12
6.3. LebesgueStieltjes Integral . . . . . . . . . . . . . . . . . . . 13
7. The Integral . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
7.1. Elementary Processes . . . . . . . . . . . . . . . . . . . . . 15
7.2. Strictly Simple and Simple Processes . . . . . . . . . . . . . . 15
8. The Stochastic Integral . . . . . . . . . . . . . . . . . . . . . 17
8.1. Integral for H L and M /
2
. . . . . . . . . . . . . . . . 17
8.2. Quadratic Variation . . . . . . . . . . . . . . . . . . . . . 19
8.3. Covariation . . . . . . . . . . . . . . . . . . . . . . . . . 22
8.4. Extension of the Integral to L
2
(M) . . . . . . . . . . . . . . . 23
8.5. Localisation . . . . . . . . . . . . . . . . . . . . . . . . . 26
8.6. Some Important Results . . . . . . . . . . . . . . . . . . . . 27
9. Semimartingales . . . . . . . . . . . . . . . . . . . . . . . . . 29
10. Relations to Sums . . . . . . . . . . . . . . . . . . . . . . . 31
10.1. The UCP topology . . . . . . . . . . . . . . . . . . . . . 31
10.2. Approximation via Riemann Sums . . . . . . . . . . . . . . . 32
11. Itos Formula . . . . . . . . . . . . . . . . . . . . . . . . . . 35
11.1. Applications of Itos Formula . . . . . . . . . . . . . . . . . 40
11.2. Exponential Martingales . . . . . . . . . . . . . . . . . . . 41
12. Levy Characterisation of Brownian Motion . . . . . . . . . . . 46
13. Time Change of Brownian Motion . . . . . . . . . . . . . . . 48
13.1. Gaussian Martingales . . . . . . . . . . . . . . . . . . . . 49
14. Girsanovs Theorem . . . . . . . . . . . . . . . . . . . . . . 51
14.1. Change of measure . . . . . . . . . . . . . . . . . . . . . 51
15. Brownian Martingale Representation Theorem . . . . . . . . . 53
16. Stochastic Dierential Equations . . . . . . . . . . . . . . . . 56
17. Relations to Second Order PDEs . . . . . . . . . . . . . . . . 61
17.1. Innitesimal Generator . . . . . . . . . . . . . . . . . . . . 61
17.2. The Dirichlet Problem . . . . . . . . . . . . . . . . . . . . 62
[ii]
Contents iii
17.3. The Cauchy Problem . . . . . . . . . . . . . . . . . . . . 64
17.4. FeynmanKac Representation . . . . . . . . . . . . . . . . . 66
18. Stochastic Filtering . . . . . . . . . . . . . . . . . . . . . . . 69
18.1. Signal Process . . . . . . . . . . . . . . . . . . . . . . . 69
18.2. Observation Process . . . . . . . . . . . . . . . . . . . . . 70
18.3. The Filtering Problem . . . . . . . . . . . . . . . . . . . . 70
18.4. Change of Measure . . . . . . . . . . . . . . . . . . . . . 70
18.5. The Unnormalised Conditional Distribution . . . . . . . . . . . 76
18.6. The Zakai Equation . . . . . . . . . . . . . . . . . . . . . 78
18.7. KushnerStratonowich Equation . . . . . . . . . . . . . . . . 86
19. Gronwalls Inequality . . . . . . . . . . . . . . . . . . . . . . 87
20. Kalman Filter . . . . . . . . . . . . . . . . . . . . . . . . . . 89
20.1. Conditional Mean . . . . . . . . . . . . . . . . . . . . . . 89
20.2. Conditional Covariance . . . . . . . . . . . . . . . . . . . . 90
21. Discontinuous Stochastic Calculus . . . . . . . . . . . . . . . 92
21.1. Compensators . . . . . . . . . . . . . . . . . . . . . . . . 92
21.2. RCLL processes revisited . . . . . . . . . . . . . . . . . . . 94
22. References . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
3. Stochastic Processes
The following notes are a summary of important denitions and results from the theory of
stochastic processes, proofs may be found in the usual books for example [Durrett, 1996].
3.1. Probability Space
Let (, T, P) be a probability space. The set of Pnull subsets of is dened by
^ := N : N A for A T, with P(A) = 0 .
The space (, T, P) is said to be complete if for A B with B T and P(B) = 0
then this implies that A T.
In addition to the probability space (, T, P), let (E, c) be a measurable space, called
the state space, which in many of the cases considered here will be (R, B), or (R
n
, B). A
random variable is a T/c measurable function X : E.
3.2. Stochastic Process
Given a probability space (, T, P) and a measurable state space (E, c), a stochastic
process is a family (X
t
)
t0
such that X
t
is an E valued random variable for each time
t 0. More formally, a map X : (R
+
, B
+
T) (R, B), where B
+
are the Borel sets
of the time space R
+
.
Denition 1. Measurable Process
The process (X
t
)
t0
is said to be measurable if the mapping (R
+
, B
+
T) (R, B) :
(t, ) X
t
() is measurable on R with respect to the product eld B(R) T.
Associated with a process is a ltration, an increasing chain of algebras i.e.
T
s
T
t
if 0 s t < .
Dene T
by
T
t0
T
t
:=
_
_
_
t0
T
t
_
_
.
If (X
t
)
t0
is a stochastic process, then the natural ltration of (X
t
)
t0
is given by
T
X
t
:= (X
s
: s t).
The process (X
t
)
t0
is said to be (T
t
)
t0
adapted, if X
t
is T
t
measurable for each t 0.
The process (X
t
)
t0
is obviously adapted with respect to the natural ltration.
[1]
Stochastic Processes 2
Denition 2. Progressively Measurable Process
A process is progressively measurable if for each t its restriction to the time interval [0, t],
is measurable with respect to B
[0,t]
T
t
, where B
[0,t]
is the Borel algebra of subsets of
[0, t].
Why on earth is this useful? Consider a noncontinuous stochastic process X
t
. From
the denition of a stochastic process for each t that X
t
T
t
. Now dene Y
t
= sup
s[0,t]
X
s
.
Is Y
s
a stochastic process? The answer is not necessarily sigma elds are only guaranteed
closed under countable unions, and an event such as
Y
s
> 1 =
_
0ss
X
s
> 1
is an uncountable union. If X were progressively measurable then this would be sucient
to imply that Y
s
is T
s
measurable. If X has suitable continuity properties, we can restrict
the unions which cause problems to be over some dense subset (say the rationals) and this
solves the problem. Hence the next theorem.
Theorem 3.3.
Every adapted right (or left) continuous, adapted process is progressively measurable.
Proof
We consider the process X restricted to the time interval [0, s]. On this interval for each
n N we dene
X
n
1
:=
2
n
1
k=0
1
(ks/2
n
,(k+1)s/2
n
]
(t)X
ks/2
n(),
X
n
2
:= 1
[0,s/2
n
)
(t)X
0
() +
2
n
k=1
1
[ks/2
n
,(k+1)s/2
n
)
(t)X
(k+1)s/2
n()
Note that X
n
1
is a left continuous process, so if X is left continuous, working pointwise
(that is, x ), the sequence X
n
1
converges to X.
But the individual summands in the denition of X
n
1
are by the adpatedness of X
clearly B
[0,s]
T
s
measurable, hence X
n
1
is also. But the convergence implies X is also;
hence X is progressively measurable.
Consideration of the sequence X
n
2
yields the same result for right continuous, adapted
processes.
The following extra information about ltrations should probably be skipped on a
rst reading, since they are likely to appear as excess baggage.
Stochastic Processes 3
Dene
t (0, ) T
t
=
0s<t
T
s
t [0, ) T
t+
=
ts<
T
s
,
whence it is clear that for each t, T
t
T
t
T
t+
.
Denition 3.2.
The family T
t
is called right continuous if
t [0, ) T
t
= T
t+
.
Denition 3.3.
A process (X
t
)
t0
is said to be bounded if there exists a universal constant K such that
for all and t 0, then [X
t
()[ < K.
Denition 3.4.
Let X = (X
t
)
t0
be a stochastic process dened on (, T, P), and let X
= (X
t
)
t0
be a
stochastic process dened on (, T, P). Then X and X
(X
t
1
A
1
, X
t
2
A
2
, . . . , X
t
n
A
n
).
Denition 3.5.
Let X and X
t
()) = 1 t 0.
Denition 3.6.
Let X and X
t
()t 0) = 1.
There is a chain of implications
indistinguishable modications same f.d.d.
The following denition provides us with a special name for a process which is indistin
guishable from the zero process. It will turn out to be important because many denitions
can only be made up to evanescence.
Denition 3.7.
A process X is evanescent if P(X
t
= 0 t) = 1.
4. Martingales
Denition 4.1.
Let X = X
t
, T
t
, t 0 be an integrable process then X is a
(i) Martingale if and only if E(X
t
[T
s
) = X
s
a.s. for 0 s t <
(ii) Supermartingale if and only if E(X
t
[T
s
) X
s
a.s. for 0 s t <
(iii) Submartingale if and only if E(X
t
[T
s
) X
s
a.s. for 0 s t <
Theorem (Kolmogorov) 4.2.
Let X = X
t
, T
t
, t 0 be an integrable process. Then dene T
t+
:=
_
>0
T
t+
and also
the partial augmentation of T by
T
t
= (T
t+
, ^). Then if t E(X
t
) is continuous there
exists an
T
t
adapted stochastic process
X =
X
t
,
T
t
, t 0 with sample paths which are
right continuous, with left limits (CADLAG) such that X and
X are modications of each
other.
Denition 4.3.
A martingale X = X
t
, T
t
, t 0 is said to be an L
2
martingale or a square integrable
martingale if E(X
2
t
) < for every t 0.
Denition 4.4.
A process X = X
t
, T
t
, t 0 is said to be L
p
bounded if and only if sup
t0
E([X
t
[
p
) < .
The space of L
2
bounded martingales is denoted by /
2
, and the subspace of continuous
L
2
bounded martingales is denoted /
c
2
.
Denition 4.5.
A process X = X
t
, T
t
, t 0 is said to be uniformly integrable if and only if
sup
t0
E
_
[X
t
[1
X
t
N
_
0 as N .
Orthogonality of Martingale Increments
A frequently used property of a martingale M is the orthogonality of increments property
which states that for a square integrable martingale M, and Y T
s
with E(Y
2
) < then
E[Y (M
t
M
s
)] = 0 for t s.
Proof
Via Cauchy Schwartz inequality E[Y (M
t
M
s
)[ < , and so
E(Y (M
t
M
s
)) = E(E(Y (M
t
M
s
)[T
s
)) = E(Y E(M
t
M
s
[T
s
)) = 0.
A typical example is Y = M
s
, whence E(M
s
(M
t
M
s
)) = 0 is obtained. A common
application is to the dierence of two squares, let t s then
E((M
t
M
s
)
2
[T
s
) =E(M
2
t
[T
s
) 2M
s
E(M
t
[T
s
) +M
2
s
=E(M
2
t
M
2
s
[T
s
) = E(M
2
t
[T
s
) M
2
s
.
[4]
Martingales 5
4.1. Stopping Times
A random variable T : [0, ) is a stopping (optional) time if and only if : T()
t T
t
.
The following theorem is included as a demonstration of checking for stopping times,
and may be skipped if desired.
Theorem 4.6.
T is a stopping time with respect to T
t+
if and only if for all t [0, ), the event T < t
if T
t
measurable.
Proof
If T is an T
t+
stopping time then for all t (0, ) the event T t is T
t+
measurable.
Thus for 1/n < t we have
_
T t
1
n
_
T
(t1/n)
+ T
t
so
T < t =
_
n=1
_
T t
1
n
_
T
t
.
To prove the converse, note that if for each t [0, ) we have that T < t T
t
,
then for each such t
_
T < t +
1
n
_
T
t+1/n
,
as a consequence of which
T t =
n=1
_
T < t +
1
n
_
n=1
T
t+1/n
= T
t+
.
Given a stochastic process X = (X
t
)
t0
, a stopped process X
T
may be dened by
X
T
() :=X
T()t
(),
T
T
:=A T : A T t T
t
.
Theorem (Optional Stopping).
Let X be a right continuous integrable, T
t
adapted process. Then the following are equiv
alent:
(i) X is a martingale.
(ii) X
T
is a martingale for all stopping times T.
(iii) E(X
T
) = E(X
0
) for all bounded stopping times T.
(iv) E(X
T
[T
S
) = X
S
for all bounded stopping times S and T such that S T. If in
addition, X is uniformly integrable then (iv) holds for all stopping times (not necessarily
bounded).
Martingales 6
The condition which is most often forgotten is that in (iii) that the stopping time T
be bounded. To see why it is necessary consider B
t
a Brownian Motion starting from zero.
Let T = inft 0 : X
t
= 1, clearly a stopping time. Equally B
t
is a martingale with
respect to the ltration generated by B itself, but it is also clear that EB
T
= 1 ,= EB
0
= 0.
Obviously in this case T < is false.
Theorem (Doobs Martingale Inequalities).
Let M = M
t
, T
t
, t 0 be a uniformly integrable martingale, and let M
:= sup
t0
[M
t
[.
Then
(i) Maximal Inequality. For > 0,
P(M
) E[[M
[1
M
<
] .
(ii) L
p
maximal inequality. For 1 < p < ,
M

p
p
p 1
M

p
.
Note that the norm used in stating the Doob L
p
inequality is dened by
M
p
= [E([M[
p
)]
1/p
.
Theorem (Martingale Convergence).
Let M = M
t
, T
t
, t 0 be a martingale.
(i) If M is L
p
bounded then M
() := lim
t
M
t
() Pa.s.
(ii) If p = 1 and M is uniformly integrable then lim
t
M
t
() = M
() in L
1
. Then
for all A L
1
(T
:= lim
t
T
t
.
(iii) If p > 1 i.e. M is L
p
bounded lim
t
M
t
= M
in L
p
.
Denition 4.7.
Let /
2
denote the set of L
2
bounded CADLAG martingales i.e. martingales M such that
sup
t0
EM
2
t
< .
Let /
c
2
denote the set of L
2
bounded CADLAG martingales which are continuous. A norm
may be dened on the space /
2
by M
2
= M

2
2
= E(M
2
).
From the conditional Jensens inequality, since f(x) = x
2
is convex,
E
_
M
2
[T
t
_
(E(M
[T
t
))
2
E
_
M
2
[T
t
_
(EM
t
)
2
.
Hence taking expectations
EM
2
t
EM
2
,
and since by martingale convergence in L
2
, we get E(M
2
t
) E(M
2
), it is clear that
E(M
2
) = sup
t0
E(M
2
t
).
Martingales 7
Theorem 4.8.
The space (/
2
, ) (up to equivalence classes dened by modications) is a Hilbert space,
with /
c
2
a closed subspace.
Proof
We prove this by showing a one to one correspondence between /
2
(the space of square
integrable martingales) and L
2
(T
)
f :(M
t
)
t0
M
lim
t
M
t
g :L
2
(T
) /
2
g :M
M
t
E(M
[T
t
)
Notice that
sup
t
EM
2
t
= M

2
2
= E(M
2
) < ,
as M
t
is a square integrable martingale. As L
2
(T
) is a Hilbert space, /
2
inherits this
structure.
To see that /
c
2
is a closed subspace of /
2
, consider a Cauchy sequence M
(n)
in
/
2
, equivalently M
(n)
is Cauchy in L
2
(T
). Hence M
(n)
converges to a limit, M
say,
in L
2
(T
). Let M
t
:= E(M
[T
t
), then
sup
t0
M
(n)
t
M
t
0, in L
2
,
that is M
(n)
M uniformly in L
2
. Hence there exists a subsequence n(k) such that
M
n(k)
M uniformly; as a uniform limit of continuous functions is continuous, M /
c
2
.
Thus /
c
2
is a closed subspace of /.
5. Basics
5.1. Local Martingales
A martingale has already been dened, but a weaker denition will prove useful for stochas
tic calculus. Note that Ill often drop references to the ltration T
t
, but this nevertheless
forms an essential part of the (local) martingale.
Just before we dive in and dene a Local Martingale, maybe we should pause and
consider the reason for considering them. The important property of local martingales
will only be seen later in the notes; and as we frequently see in this subject it is one of
stability that is, they are a class of objects which are closed under an operation, in this case
under the stochastic integral an integral of a previsible process with a local martingale
integrator is a local martingale.
Denition 5.1.
M = M
t
, T
t
, 0 t is a local martingale if and only if there exists a sequence of
stopping times T
n
tending to innity such that M
T
n
are martingales for all n. The space
of local martingales is denotes /
loc
, and the subspace of continuous local martingales is
denotes /
c
loc
.
Recall that a martingale (X
t
)
t0
is said to be bounded if there exists a universal
constant K such that for all and t 0, then [X
t
()[ < K.
Theorem 5.2.
Every bounded local martingale is a martingale.
Proof
Let T
n
be a sequence of stopping times as in the denition of a local martingale. This
sequence tends to innity, so pointwise X
T
n
t
() X
t
(). Using the conditional form of the
dominated convergence theorem (using the constant bound as the dominating function),
for t s 0
lim
n
E(X
T
n
t
[T
s
) = E(X
t
[T
s
).
But as X
T
n
is a (genuine) martingale, E(X
T
n
t
[T
s
) = X
T
n
s
= X
T
n
s
; so
E(X
t
[T
s
) = lim
n
E(X
T
n
t
[T
s
) = lim
n
X
T
n
s
= X
s
.
Hence X
t
is a genuine martingale.
Proposition 5.3.
The following are equivalent
(i) M = M
t
, T
t
, 0 t is a continuous martingale.
(ii) M = M
t
, T
t
, 0 t is a continuous local martingale and for all t 0, the set
M
T
: T a stopping time, T t is uniformly integrable.
Proof
(i) (ii) By optional stopping theorem, if T t then M
T
= E(M
t
[T
T
) hence the set is
uniformly integrable.
[8]
Basics 9
(ii) (i)It is required to prove that E(M
0
) = E(M
T
) for any bounded stopping time T.
Then by local martingale property for any n,
E(M
0
) = E(M
TT
n
),
uniform integrability then implies that
lim
n
E(M
TT
n
) = E(M
T
).
5.2. Local Martingales which are not Martingales
There do exist local martingales which are not themselves martingales. The following is
an example Let B
t
be a d dimensional Brownian Motion starting from x. It can be shown
using Itos formula that a harmonic function of a Brownian motion is a local martingale
(this is on the example sheet). From standard PDE theory it is known that for d 3, the
function
f(x) =
1
[x[
d2
is a harmonic function, hence X
t
= 1/[B
t
[
d2
is a local martingale. Now consider the L
p
norm of this local martingale
E
x
[X
t
[
p
=
_
1
(2t)
d/2
exp
_
[y x[
2
2t
_
[y[
(d2)p
dy.
Consider when this integral converges. There are no divergence problems for [y[ large, the
potential problem lies in the vicinity of the origin. Here the term
1
(2t)
d/2
exp
_
[y x[
2
2t
_
is bounded, so we only need to consider the remainder of the integrand integrated over a
ball of unit radius about the origin which is bounded by
C
_
B(0,1)
[y[
(d2)p
dy,
for some constant C, which on tranformation into polar coordinates yields a bound of the
form
C
_
1
0
r
(d2)p
r
d1
dr,
with C
_
_
_
n()
k=1
A
t
k
t
A
t
k1
t
: 0 = t
0
t
1
t
n
= t
_
_
_
.
An alternative denition is given by
V
0
t
:= lim
n
n()
A
k2
n
t
A
(k1)2
n
t
.
These can be shown to be equivalent (for CADLAG processes), since trivially (use the
dyadic partition), V
0
t
V
t
. It is also possible to show that V
0
t
V
t
for the total variation
of a CADLAG process.
Denition 6.1.
A process A is said to have nite variation if the associated variation process V is nite
(i.e. if for every t and every , [V
t
()[ < .
6.1. Why we need a Stochastic Integral
Before delving into the depths of the integral its worth stepping back for a moment to see
why the ordinary integral cannot be used on a path at a time basis (i.e. separately for
each ). Suppose we were to do this i.e. set
I
t
(X) =
_
t
0
X
s
()dM
s
(),
for M /
c
2
; but for an interesting martingale (i.e. one which isnt zero a.s.), the total
variation is not nite, even on a bounded interval like [0, T]. Thus the LebesgueStieltjes
integral denition isnt valid in this case. To generalise we shall see that the quadratic
variation is actually the right variation to use (higher variations turn out to be zero and
lower ones innite, which is easy to prove by considering the variation expressed as the
limit of a sum and factoring it by a maximum multiplies by the quadratic variation, the
rst term of which tends to zero by continuity). But to start, we shall consider integrating
a previsible process H
t
with an integrator which is an increasing nite variation process.
First we shall prove that a continuous local martingale of nite variation is zero.
[11]
Total Variation and the Stieltjes Integral 12
Proposition 6.2.
If M is a continuous local martingale of nite variation, starting from zero then M is
identically zero.
Proof
Let V be the variation process of M. This V is a continuous, adapted process. Now dene a
sequence of stopping times S
n
as the rst time V exceeds n, i.e. S
n
:= inf
t
t 0 : V
t
n.
Then the martingale M
S
n
is of bounded variation. It therefore suces to prove the result
for a bounded, continuous martingale M of bounded variation.
Fix t 0 and let 0 = t
0
, t
1
, . . . , t
N
= t be a partition of [0, t]. Then since M
0
= 0 it is
clear that, M
2
t
=
N
k=1
_
M
2
t
k
M
2
t
k1
_
. Then via orthogonality of martingale increments
E(M
2
t
) =E
_
N
k=1
_
M
t
k
M
t
k1
_
2
_
E
_
V
t
sup
k
M
t
k
M
t
k1
_
The integrand is bounded by n
2
(from denition of the stopping time S
n
), hence the
expectation converges to zero as the modulus of the partition tends to zero by the bounded
convergence theorem. Hence M 0.
6.2. Previsibility
The term previsible has crept into the discussion earlier. Now is the time for a proper
denition.
Denition 6.3.
The previsible (or predictable) eld T is the eld on R
+
generated by the processes
(X
t
)
t0
, adapted to T
t
, with left continuous paths on (0, ).
Remark
The same eld is generated by left continuous, right limits processes (i.e. c`agl`ad pro
cesses) which are adapted to T
t
, or indeed continuous processes (X
t
)
t0
which are adapted
to T
t
. It is gnerated by sets of the form A(s, t] where A T
s
. It should be noted that
c`adl`ag processes generate the optional eld which is usually dierent.
Theorem 6.4.
The previsible eldis also generated by the collection of random sets A 0 where
A T
0
and A(s, t] where A T
s
.
Proof
Let the eld generated by the above collection of sets be denotes T
. We shall show
T = T
k
X
k/2
n1
(k/2
n
,(k+1)/2
n
]
(t)
It is clear that X
n
T
measurable, thus T T
. Conversely
Total Variation and the Stieltjes Integral 13
consider the indicator function of A (s, t] this can be written as 1
[0,t
A
]\[0,s
A
]
, where
s
A
() = s for A and + otherwise. These indicator functions are adapated and left
continuous, hence T
T.
Denition 6.5.
A process (X
t
)
t0
is said to be previsible, if the mapping (t, ) X
t
() is measurable
with respect to the previsible eld T.
6.3. LebesgueStieltjes Integral
[In the lecture notes for this course, the LebesgueStieltjes integral is considered rst for
functions A and H; here I consider processes on a pathwise basis.]
Let A be an increasing cadlag process. This induces a Borel measure dA on (0, )
such that
dA((s, t])() = A
t
() A
s
().
Let H be a previsible process (as dened above). The LebesgueStieltjes integral of H is
dened with respect to an increasing process A by
(H A)
t
() =
_
t
0
H
s
()dA
s
(),
whenever H 0 or ([H[ A)
t
< .
As a notational aside, we shall write
(H A)
t
_
t
0
HdX,
and later on we shall use
d(H X) HdX.
This denition may be extended to integrator of nite variation which are not increas
ing, by decomposing the process A of nite variation into a dierence of two increasing
processes, so A = A
+
A
, where A
)
t
(),
whenever ([H[ V )
t
< .
There are no really new concepts of the integral in the foregoing; it is basically the
LebesgueStieltjes integral eextended from functions H(t) to processes in a pathwise fashion
(thats why has been included in those denitions as a reminder).
Theorem 6.6.
If X is a nonnegative continuous local martingale and E(X
0
) < then X
t
is a super
martingale. If additionally X has constant mean, i.e. E(X
t
) = E(X
0
) for all t then X
t
is a
martingale.
Total Variation and the Stieltjes Integral 14
Proof
As X
t
is a continuous local martingale there is a sequence of stopping times T
n
such
that X
T
n
is a genuine martingale. From this martingale property
E(X
T
n
t
[T
s
) = X
T
n
s
.
As X
t
0 we can apply the conditional form of Fatous lemma, so
E(X
t
[T
s
) = E(liminf
n
X
T
n
t
[T
s
) liminf
n
E(X
T
n
t
[T
s
) = liminf
n
X
T
n
s
= X
s
.
Hence E(X
t
[T
s
) X
s
, so X
t
is a supermartingale.
Given the constant mean property E(X
t
) = E(X
s
). Let
A
n
:= : X
s
E(X
t
[T
s
) > 1/n,
so
A :=
_
n=1
A
n
= : X
s
E(X
t
[T
s
) > 0.
Consider P(A) = P(
n=1
A
n
)
n=1
P(A
n
). Suppose for some n, P(A
n
) > , then note
that
A
n
: X
s
E(X
t
[T
s
) > 1/n
/A
n
: X
s
E(X
t
[T
s
) 0
Hence
X
s
E(X
t
[T
s
)
1
n
1
A
n
,
taking expectations yields
E(X
s
) E(X
t
) >
n
,
but by the constant mean property the left hand side is zero; hence a contradiction, thus
all the P(A
n
) are zero, so
X
s
= E(X
t
[T
s
) a.s.
7. The Integral
We would like eventually to extend the denition of the integral to integrands which are
previsible processes and integrators which are semimartingales (to be dened later in these
notes). In fact in these notes well only get as far as continuous semimartingales; but it is
possible to go the whole way and dene the integral of a previsible process with respect to
a general semimartingale; but some extra problems are thrown up on the way, in particular
as regards the construction of the quadratic variation process of a discontinuous process.
Various special classes of process will be needed in the sequel and these are all dened
here for convenience. Naturally with terms like elementary and simple occurring many
books have dierent names for the same concepts so beware!
7.1. Elementary Processes
An elementary process H
t
() is one of the form
H
t
() = Z()1
(S(),T()]
(t),
where S, T are stopping times, S T , and Z is a bounded T
S
measurable random
variable.
Such a process is the simplest nontrivial example of a previsible process. Lets prove
that it is previsible:
H is clearly a left continuous process, so we need only show that it is adapted. It can
be considered as the pointwise limit of a sequence of right continuous processes
H
n
(t) = lim
n
Z1
[S
n
,T
n
)
, S
n
= S +
1
n
, T
n
= T +
1
n
.
So it is sucient to show that Z1
[U,V )
is adapted when U and V are stopping times which
satisfy U V , and Z is a bounded T
U
measurable function. Let B be a borel set of R,
then the event
Z1
[U,V )
(t) B = [Z B U t] V > t.
By the denition of U as a stopping time and hence the denition of T
U
, the event enclosed
by square brackets is in T
t
, and since V is a stopping time V > t = /V t is also
in T
t
; hence Z1
[U,V )
is adapted.
7.2. Strictly Simple and Simple Processes
A process H is strictly simple (H L
) if there exist 0 t
0
t
n
< and uniformly
bounded T
t
k
measurable random variables Z
k
such that
H = H
0
()1
0
(t)
n1
k=0
Z
k
()1
(t
k
,t
k+1
](t)
.
[15]
The Integral 16
This can be extended to H is a simple processes (H L), if there exists a sequence
of stopping times 0 T
0
T
k
, and Z
k
uniformly bounded T
T
k
measurable
random variables such that
H = H
0
()1
0
(t) +
k=0
Z
k
1
(T
k
,T
k+1
]
.
Similarly a simple process is also a previsible process. The fundamental result will
follow from the fact that the algebra generated by the simple processes is exactly the
previsible algebra. We shall see the application of this after the next section.
8. The Stochastic Integral
As has been hinted at earlier the stochastic integral must be built up in stages, and to
start with we shall consider integrators which are L
2
bounded martingales, and integrands
which are simple processes.
8.1. Integral for H L and M /
2
For a simple process H L, and M an L
2
bounded martingale then the integral may be
dened by the martingale transform (c.f. discrete martingale theory)
_
t
0
H
s
dM
s
= (H M)
t
:=
k=0
Z
k
_
M
T
k+1
t
M
T
k
t
_
Proposition 8.1.
If H is a simple process, M a L
2
bounded martingale, and T a stopping time. Then
(i) (H M)
T
= (H1
(0,T]
) M = H (M
T
).
(ii) (H M) /
2
.
(iii) E[(H M)
2
] =
k=0
E[Z
2
k
(M
2
T
k+1
M
2
T
k
)] H
2
E(M
2
).
Proof
Part (i)
As H L we can write
H =
k=0
Z
k
1
(T
k
,T
k+1
]
,
for T
k
stopping times, and Z
k
an T
T
k
measurable bounded random variable. By our de
nition for M /
2
, we have
(H M)
t
=
k=0
Z
k
_
M
T
k+1
t
M
T
k
t
_
,
and so, for T a general stopping time consider (H M)
T
t
= (H M)
Tt
and so
(H M)
T
t
=
k=0
Z
k
_
M
T
k+1
Tt
M
T
k
Tt
_
.
Similar computations can be performed for (H M
T
), noting that M
T
t
= M
Tt
and for
(H1
(0,T]
M) yielding the same result in both cases. Hence
(H M)
T
= (H1
(0,T]
M) = (H M
T
).
[17]
The Stochastic Integral 18
Part (ii)
To prove this result, rst we shall establish it for an elementary process H c, and then
extend to L by linearity. Suppose
H = Z1
(R,S]
,
where R and S are stopping times and Z is a bounded T
R
measurable random variable.
Let T be an arbitrary stopping time. We shall prove that
E((H M)
T
) = E((H M)
0
) ,
and hence via optional stopping conclude that (H M)
t
is a martingale.
Note that
(H M)
= Z (M
S
M
R
) ,
and hence as M is a martingale, and Z is T
R
measurable we obtain
E(H M)
=E(E(Z (M
S
M
R
)) [T
R
) = E(ZE((M
S
M
R
) [T
R
))
=0.
Via part (i) note that E(H M)
T
= E(H M
T
), so
E(H M)
T
= E(H M
T
)
= 0.
Thus (H M)
t
is a martingale by optional stopping theorem. By linearity, this result
extends to H a simple process (i.e. H L).
Part (iii)
We wish to prove that (H M) is an L
2
bounded martingale. We again start by considering
H c, an elementary process, i.e.
H = Z1
(R,S]
,
where as before R and S are stopping times, and Z is a bounded T
R
measurable random
variable.
E
_
(H M)
2
_
=E
_
Z
2
(M
S
M
R
)
2
_
,
=E
_
Z
2
E
_
(M
S
M
R
)
2
[T
R
__
,
where Z
2
is removed from the conditional expectation since it is and T
R
measurable
random variable. Using the same argument as used in the orthogonality of martingale
increments proof,
E
_
(H M)
2
_
= E
_
Z
2
E
_
(M
2
S
M
2
R
)[T
R
__
= E
_
Z
2
_
M
2
S
M
2
R
__
.
As M is an L
2
bounded martingale and Z is a bounded process,
E
_
(H M)
2
_
sup
2[Z()[
2
E
_
M
2
_
.
The Stochastic Integral 19
so (H M) is an L
2
bounded martingale; so together with part (ii), (H M) /
2
.
To extend this to simple processes is similar, but requires a little care In general the
orthogonality of increments arguments extends to the case where only nitely many of the
Z
k
in the denition of the simple process H are non zero. Let K be the largest k such that
Z
k
, 0.
E
_
(H M)
2
_
=
K
k=0
E
_
Z
2
k
_
M
2
T
k+1
M
2
T
k
__
,
which can be bounded as
E
_
(H M)
2
_
H

2
E
_
K
k=0
_
M
2
T
k+1
M
2
T
k
_
_
H

2
E
_
M
2
T
K+1
M
2
T
0
_
H

2
EM
2
,
since we require T
0
= 0, and M /
2
, so the nal bound is obtained via the L
2
martingale
convergence theorem.
Now extend this to the case of an innite sum; let n m, we have that
(H M)
T
m
(H M)
T
n
= (H1
(T
n
,T
m
]
M),
applying the result just proven for nite sums to the right hand side yields
_
_
(H M)
T
m
(H M)
T
n
_
_
2
2
=
m1
k=n
E
_
Z
2
k
_
M
2
T
k+1
M
2
T
k
__
H

2
2
E
_
M
2
M
2
T
n
_
.
But by the L
2
martingale convergence theorem the right hand side of this bound tends to
zero as n ; hence (H M)
T
n
converges in /
2
and the limit must be the pointwise
limit (H M). Let n = 0 and m and the result of part (iii) is obtained.
8.2. Quadratic Variation
We mentioned earlier that the total variation is the variation which is used by the usual
LebesgueStieltjes integral, and that this cannot be used for dening a stochastic integral,
since any continuous local martingale of nite variation is indistinguishable from zero. We
are now going to look at a variation which will prove fundamental for the construction of
the integral. All the denitions as given here arent based on the partition construction.
This is because I shall follow Dellacherie and Meyer and show that the other denitions
are equivalent by using the stochastic integral.
Theorem 8.2.
The quadratic variation process M)
t
of a continuous L
2
integrable martingale M is
the unique process A
t
starting from zero such that M
2
t
A
t
is a uniformly integrable
martingale.
The Stochastic Integral 20
Proof
For each n dene stopping times
S
n
0
:= 0, S
n
k+1
:= inf
_
t > S
n
k
:
M
t
M
S
n
k
> 2
n
_
for k 0
Dene
T
n
k
:= S
n
k
t
Then
M
2
t
=
k1
_
M
2
tS
n
k
M
2
tS
n
k1
_
=
k1
_
M
2
T
n
k
M
2
T
n
k1
_
=2
k1
M
T
n
k1
_
M
T
n
k
M
T
n
k1
_
+
k1
_
M
T
n
k
M
T
n
k1
_
2
()
Now dene H
n
to be the simple process given by
H
n
:=
k1
M
S
n
k1
1
(S
n
k1
,S
n
k
]
.
We can then think of the rst term in the decomposition () as (H
n
M). Now dene
A
n
t
:=
k1
_
M
T
n
k
M
T
n
k1
_
2
,
so the expression () becomes
M
2
t
= 2(H
n
M)
t
+A
n
t
. ()
From the construction of the stopping times S
n
k
we have the following properties
H
n
H
n+1

=sup
t
[H
n
t
H
n+1
t
[ 2
(n+1)
H
n
H
n+m

=sup
t
[H
n
t
H
n+m
t
[ 2
(n+1)
for all m 1
H
n
M
=sup
t
[H
n
t
M
t
[ 2
n
Let J
n
() be the set of all stopping times S
n
k
() i.e.
J
n
() := S
n
k
() : k 0.
Clearly J
n
() J
n+1
(). Now for any m 1, using Proposition 8.1(iii) the following
result holds
E
_
_
(H
n
M) (H
n+m
M)
_
=E
_
___
H
n
H
n+m
_
M
_
2
_
H
n
H
n+m

2
E(M
2
_
2
(n+1)
_
2
E(M
2
).
The Stochastic Integral 21
Thus (H
n
M)
); hence by
completeness of the Hilbert Space it converges to a limit in the same space. As (H
n
M)
is a continuous martingale for each n, so is the the limit N say. By Doobs L
2
inequality
applied to the continuous martingale (H
n
M) N,
E
_
sup
t0
[(H
n
M) N[
2
_
4E
_
[(H M) N]
2
n
0.
Hence (H
n
M) converges to N uniformly a.s.. From the relation () we see that as a
consequence of this, the process A
n
converges uniformly a.s. to a process A, where
M
2
t
= 2N
t
+A
t
.
Now we must check that this limit process A is increasing. Clearly A
n
(S
n
k
) A
n
(S
n
k+1
),
and since J
n
() J
n+1
(), it is also true that A(S
n
k
) A(S
n
k+1
) for all n and k, and
so A is certainly increasing on the closure of J() :=
n
J
n
(). However if I is an open
interval in the complement of J, then no stopping time S
n
k
lies in this interval, so M must
be constant throughout I, so the same is true for the process A. Hence the process A
is continuous, increasing, and null at zero; such that M
2
t
A
t
= 2N
t
, where N
t
is a UI
martingale (since it is L
2
bounded). Thus we have established the existence result. It only
remains to consider uniqueness.
Uniqueness follows from the result that a continuous local martingale of nite variation
is everywhere zero. Suppose the process A in the above denition were not unique. That is
suppose that also for some B
t
continuous increasing from zero, M
2
t
B
t
is a UI martingale.
Then as M
2
t
A
t
is also a UI martingale by subtracting these two equations we get that
A
t
B
t
is a UI martingale, null at zero. It clearly must have nite variation, and hence be
zero.
The following corollary will be needed to prove the integration by parts formula, and
can be skipped on a rst reading; however it is clearer to place it here, since this avoids
having to redene the notation.
Corollary 8.3.
Let M be a bounded continuous martingale, starting from zero. Then
M
2
t
= 2
_
t
0
M
s
dM
s
+M)
t
.
Proof
In the construction of the quadratic variation process the quadratic variation was con
structed as the uniform limit in L
2
of processes A
n
t
such that
A
n
t
= M
2
t
2(H
n
M)
t
,
where each H
n
was a bounded previsible process, such that
sup
t
[H
n
t
M[ 2
n
,
The Stochastic Integral 22
and hence H
n
M in L
2
(M), so the martingales (H
n
M) converge to (M M) uniformly
in L
2
, hence it follows immediately that
M
2
t
= 2
_
t
0
M
s
dM
s
+M)
t
,
Theorem 8.4.
The quadratic variation process M)
t
of a continuous local martingale M is the unique
increasing process A, starting from zero such that M
2
A is a local martingale.
Proof
We shall use a localisation technique to extend the denition of quadratic variation from
L
2
bounded martingales to general local martingales.
The mysterious seeming technique of localisation isnt really that complex to under
stand. The idea is that it enables us to extend a denition which applies for X widgets to
one valid for local X widgets. It achieves this by using a sequence of stopping times which
reduce the local X widgets to X widgets ; the original denition can then be applied to
the stopped version of the X widget. We only need to check that we can sew up the pieces
without any holes i.e. that our denition is independent of the choice of stopping times!
Let T
n
= inft : [M
t
[ > n, dene a sequence of stopping times. Now dene
M)
t
:= M
T
n
) for 0 t T
n
To check the consistency of this denition note that
M
T
n
)
T
n1
= M
T
n1
)
and since the sequence of stopping times T
n
, we see that M) is dened for all t.
Uniqueness follows from the result that any nite variation continuous local martingale
starting from zero is identically zero.
The quadratic variation turns out to be the right sort of variation to consider for
a martingale; since we have already shown that all but the zero martingale have innite
total variation; and it can be shown that the higher order variations of a martingale are
zero a.s.. Note that the denition given is for a continuous local martingale; we shall
see later how to extend this to a continuous semimartingale.
8.3. Covariation
From the denition of the quadratic variation of a local martingale we can dene the covari
ation of two local martingales N and M which are locally L
2
bounded via the polarisation
identity
M, N) :=
M +N) M N)
4
.
We need to generalise this slightly, since the above denition required the quadratic
variation terms to be nite. We can prove the following theorem in a straightforward
manner using the denition of quadratic variation above, and this will motivate the general
denition of the covariation process.
The Stochastic Integral 23
Theorem 8.5.
For M and N two local martingales which are locally L
2
bounded then there exists a unique
nite variation process A starting from zero such that MN A is a local martingale. This
process A is the covariation of M and N.
This theorem is turned round to give the usual denition of the covariation process of
two continuous local martingales as:
Denition 8.6.
For two continuous local martingales N and M, there exists a unique nite variation
process A, such that MN A is a local martingale. The covariance process of N and M
is dened as this process A.
It can readily be veried that the covariation process can be regarded as a symmet
ric bilinear form on the space of local martingales, i.e. for L,M and N continuous local
martingales
M +N, L) =M, L) +N, L),
M, N) =N, M),
M, N) =M, N), R.
8.4. Extension of the Integral to L
2
(M)
We have previously dened the integral for H a simple process (in L), and M /
c
2
, and
we have noted that (H M) is itself in /
2
. Hence
E
_
(H M)
2
_
= E
_
Z
2
i1
_
M
T
i
M
T
i1
_
2
_
Recall that for M /
2
, then M
2
M) is a uniformly integrable martingale. Hence for
S and T stopping times such that S T, then
E
_
(M
T
M
S
)
2
[T
S
_
= E(M
2
T
M
2
S
[T
S
) = E(M)
T
M)
S
[T
S
).
So summing we obtain
E
_
(H M)
2
_
=E
Z
2
i1
_
M)
T
i
M)
T
i1
_
,
=E
_
(H
2
M))
_
.
In the light of this, we dene a seminorm H
M
via
H
M
=
_
E
_
(H
2
M))
_
_
1/2
=
_
E
__
0
H
2
s
dM)
s
__
1/2
.
The space L
2
(M) is then dened as the subspace of the previsible processes, where this
seminorm is nite, i.e.
L
2
(M) := previsible processes H such that H
M
< .
The Stochastic Integral 24
However we would actually like to be able to treat this as a Hilbert space, and there
remains a problem, namely that if X L
2
(M) and X
M
= 0, this doesnt imply that X
is the zero process. Thus we follow the usual route of dening an equivalence relation via
X Y if and only if X Y 
M
= 0. We now dene
L
2
(M) := equivalence classes of previsible processes H such that H
M
< ,
and this is a Hilbert space with norm  
M
(it can be seen that it is a Hilbert space by
considering it as suitable L
2
space).
This establishes an isometry (called the Ito isometry) between the spaces L
2
(M) L
and L
2
(T
) given by
I :L
2
(M) L L
2
(T
)
I :H (H M)
in L
2
(T
) .
We shall now use this Ito isometry to extend the denition of the stochastic integral
from L (the class of simple processes) to the whole of L
2
(M). Roughly speaking we shall
approximate an element of L
2
(M) via a sequence of simple processes converging to it; just
as in the construction of the Lebesgue Integral. In doing this, we shall use the Monotone
Class Theorem.
Recall that in the conventional construction of the Lebesgue integration, and proof of
the elementary results the following standard machine is repeatedly invoked. To prove a
linear result for all h L
1
(S, , ), proceed in the following way:
(i) Show the result is true for h an indicator function.
(ii) Show that by linearity the result extends to all positive step functions.
(iii) Use the Monotone convergence theorem to see that if h
n
h, where the h
n
are
step functions, then the result must also be true for h a nonnegative, measurable
function.
(iv) Write h = h
+
h
where both h
+
and h
i=2
H
(i1)/n
_
i 1
n
,
i
n
_
,
and if H
t
is adapted to T
t
then H
(i1)/n
is a bounded element of T
(i1)/n
.
We can now apply the Monotone Class Theorem to the vector space H of processes
with a time parameter in (0, ), regarded as maps from (0, ) R. Then if this
vector space contains all the simple processes i.e. L H, then H contains every bounded
previsible process on (0, ).
Assembling the Pieces
Since I is an isometry it has a unique extension to the closure of
 = L
2
(M) L,
in L
2
(M). By the application of the monotone class lemma to H = , and the system
of simple processes. We see that  must contain every bounded previsible process; hence
 = L
2
(M). Thus the Ito Isometry extends to a map from L
2
(M) to L
2
(T
).
Let us look at this result more informally. For a previsible H L
2
(M), because of the
density of L in L
2
(M), we can nd a sequence of simple processes H
n
which converges to
H, as n . We then consider I(H) as the limit of the I(H
n
). To verify that this limit
is unique, suppose that H
n
H as n also, where H
n
L. Note that H
n
H
n
L.
Also H
n
H
n
0 and so ((H
n
H
n
) M) 0, and hence by the Ito isometry the limits
lim
n
(H
n
M) and lim
n
(H
n
M) coincide.
The following result is essential in order to extend the integral to continuous local
martingales.
The Stochastic Integral 26
Proposition 8.8.
For M /
2
, for any H L
2
(M) and for any stopping time T then
(H M)
T
= (H1
(0,T]
M) = (H M
T
).
Proof
Consider the following linear maps in turn
f
1
:L
2
(T
) L
2
(T
)
f
1
:Y E(Y [T
T
)
This map is a contraction on L
2
(T
[T
T
)
2
E(Y
2
[T
T
),
and taking expectations yields
E(Y [T
T
)
2
2
= E
_
E(Y
[T
T
)
2
_
E
_
E(Y
2
[T
T
)
_
= E(Y
2
) = Y 
2
2
.
Hence f
1
is a contraction on L
2
(T
). Now
f
2
:L
2
(M) L
2
(M)
f
2
:H H1
(0,T]
Clearly from the denition of  
M
, and from the fact that the quadratic variation process
is increasing
_
_
H1
(0,T]
_
_
M
=
_
0
H
2
s
1
(0,T]
dM)
s
=
_
T
0
H
2
s
dM)
s
_
0
H
2
s
dM)
s
= H
M
.
Hence f
2
is a contraction on L
2
(M). Hence if I denotes the Ito isometry then f
1
I and
I f
2
are also contractions from L
2
(M) to L
2
(T
)).
Now introduce I
(T)
, the stochastic integral map associated with M
T
, i.e.
I
(T)
(H) (H M
T
)
.
Note that
I
(T)
(H)
2
= H
M
T H
M
.
We have previously shown that the maps f
1
I and I f
2
and H I
(T)
(H) agree on
the space of simple processes by direct calculation. We note that L is dense in L
2
(M)
(from application of Monotone Class Lemma to the simple processes). Hence from the
three bounds above the three maps agree on L
2
(M).
The Stochastic Integral 27
8.5. Localisation
Weve already met the idea of localisation in extending the denition of quadratic variation
from L
2
bounded continuous martingales to continuous local martingales. In this context
a previsible process H
t
t0
, is locally previsible if there exists a sequence of stopping
times T
n
such that for all n H1
(0,T
n
]
is a previsible process. Fairly obviously every
previsible process has this property. However if in addition we want the process H to be
locally bounded we need the condition that there exists a sequence T
n
of stopping times,
tending to innity such that H1
(0,T
n
]
is uniformly bounded for each n.
For the integrator (a martingale of integrable variation say), the localisation is to a
local martingale, that is one which has a sequence of stopping times T
n
such that
for all n, X
T
n
is a genuine martingale.
If we can prove a result like
(H X)
T
= (H1
(0,T]
X
T
)
for H and X in their original (i.e. nonlocalised classes) then it is possible to extend the
denition of (H X) to the local classes.
Note rstly that for H and X local, and T
n
a reducing sequence
1
of stopping times
for both H and X then we see that (H1
(0,T]
X
T
) is dened in the existing fashion. Also
note that if T = T
n1
we can check consistency
(H1
(0,T
n
]
X
T
n
)
T
n1
= (H X)
T
n1
= (H1
(0,T
n1
]
X
T
n1
).
Thus it is consistent to dene (H X)
t
on t [0, ) via
(H X)
T
n
= (H1
(0,T
n
]
X
T
n
), n.
We must check that this is well dened, viz if we choose another regularising sequence S
n
,
we get the same denition of (H X). To see this note:
(H1
(0,T
n
]
X
T
n
)
S
n
= (H1
(0,T
n
S
n
]
X
T
n
S
n
) = (H1
(0,S
n
]
X
S
n
)
T
n
,
hence the denition of (H X)
t
is the same if constructed from the regularising sequence
S
n
as if constructed via T
n
.
8.6. Some Important Results
We can now extend most of our results to stochastic integrals of a previsible process H
with respect to a continuous local martingale M. In fact in these notes we will never
drop the continuity requirement. It can be done; but it requires considerably more work,
especially with regard to the denition of the quadratic variation process.
1
The reducing sequence is the sequence of stopping times tending to innity which makes the local
version of the object into the nonlocal version. We can nd one such sequence, because if say {T
n
}
reduces H and {S
n
} reduces X then T
n
S
n
reduces both H and X.
The Stochastic Integral 28
Theorem 8.9.
Let H be a locally bounded previsible process, and M a continuous local martingale. Let
T be an arbitrary stopping time. Then:
(i) (H M)
T
= (H1
(0,T]
M) = (H M
T
)
(ii) (H M) is a continuous local martingale
(iii) H M) = H
2
M)
(iv) H (K M) = (HK) M
Proof
The proof of parts (i) and (ii) follows from the result used in the localisation that:
(H M)
T
= (H1
(0,T]
M) = (H M
T
)
for H bounded previsible process in L
2
(M) and M an L
2
bounded martingale. Using this
result it suces to prove (iii) and (iv) where M, H and K are uniformly bounded (via
localisation).
Part (iii)
E
_
(H M)
2
T
=E
__
H1
(0,T]
M
_
M)
2
=E
_
_
H1
(0,T]
M)
_
2
_
=E
__
H
2
M)
_
T
_
=E
__
H
2
K M)
_
=E
__
H
2
(K
2
M))
_
=E
__
(HK)
2
M)
_
=E
_
((HK) M)
2
_
Also note the following bound
E
__
HK)
2
M)
_
min
_
H
2
K
2
M
, H
2
M
K
2
_
.
9. Semimartingales
I mentioned at the start of these notes that the most general form of the stochastic integral
would have a previsible process as the integrand and a semimartingale as an integrator.
Now its time to extend the denition of the Ito integral to the case of semimartingale
integrators.
Denition 9.1.
A process X is a semimartingale if X is an adapted CADLAG process which has a decom
position
X = X
0
+M +A,
where M is a local martingale, null at zero and A is a process null at zero, with paths of
nite variation.
Note that the decomposition is not necessarily unique as there exist martingales which
have nite variation. To remove many of these diculties we shall impose a continuity
condition, since under this most of our problems will vanish.
Denition 9.2.
A continuous semimartingale is a process (X
t
)
t0
which has a DoobMeyer decomposition
X = X
0
+M +A,
where X
0
is T
0
measurable, M
0
= A
0
= 0, M
t
is a continuous local martingale and A
t
is
a continuous adapted process of nite variation.
Theorem 9.3.
The DoobMeyer decomposition in the denition of a continuous semimartingale is unique.
Proof
Let another such decomposition be
X = X
0
+M
+A
,
where M
M = A
A,
by the rst equality, N is the dierence of two continuous local martingales, and hence is
itself a continuous local martingale; and by the second inequality it has nite variation.
Hence by an earlier proposition (5.2) it must be zero. Hence M
= M and A
= A.
We dene the quadratic variation of the continuous semimartingale as that of the
continuous local martingale part i.e. for X = X
0
+M +A,
X) := M).
These denitions can be made to look natural by considering the quadratic variation dened in terms
of a sum of squared increments; but following this approach, these are result which are proved later
using the Ito integral, since this provided a better approach to the discontinuous theory.
[29]
Semimartingales 30
Similarly if Y +Y
0
+N +B is another semimartingale, where B is nite variation and N
is a continuous local martingale, we dene
X, Y ) := M, N).
We can extend the denition of the stochastic integral to continuous semimartingale
integrators by dening
(H X) := (H M) + (H A),
where the rst integral is a stochastic integral as dened earlier and the second is a
LebesgueStieltjes integral (as the integrator is a process of nite variation).
10. Relations to Sums
This section is optional; and is included to bring together the two approaches to the
constructions involved in the stochastic integral.
For example the quadratic variation of a process can either be dened in terms of
martingale properties, or alternatively in terms of sums of squares of increments.
10.1. The UCP topology
We shall meet the notion of convergence uniformly on compacts in probability when con
sidering stochastic integrals as limits of sums, so it makes sense to review this topology
here.
Denition 10.1.
A sequence H
n
n1
converges to a process H uniformly on compacts in probability (ab
breviated u.c.p.) if for each t > 0,
sup
0st
[H
n
s
H
s
[ 0 in probability.
At rst sight this may seem to be quite an esoteric denition; in fact it is a natural
extension of convergence in probability to processes. It would also appear to be quite
dicult to handle, however Doobs martingale inequalities provide the key to handling it.
Let
H
t
= sup
0st
[H
s
[,
then for Y
n
a CADLAG process, Y
n
converges to Y u.c.p. i (Y
n
Y )
converges to
zero in probability for each t 0. Thus to prove that a sequence converges u.c.p. it often
suces to apply Doobs inequality to prove that the supremum converges to zero in L
2
,
whence it must converge to zero in probability, whence u.c.p. convergence follows.
The space of CADLAG processes with u.c.p. topology is in fact metrizable, a compat
ible metric is given by
d(X, Y ) =
n=1
1
2
n
E(min(1, (X Y )
n
)) ,
for X and Y CADLAG processes. The metric space can also be shown to be complete. For
details see Protter.
[31]
Relations to Sums 32
Since we have just met a new kind of convergence, it is helpful to recall the other usual
types of convergence on a probability space. For convenience here are the usual denitions:
Pointwise
A sequence of random variables X
n
converges to X pointwise if for all not in some null
set,
X
n
() X().
Probability
A sequence of r.v.s X
n
converges to X in probability, if for any > 0,
P([X
n
X[ > ) 0, as n .
L
p
convergence
A sequence of random variables X
n
converges to X in L
p
, if
E[X
n
X[
p
0, as n .
It is trivial to see that pointwise convergence implies convergence in probability. It is
also true that L
p
convergence implies convergence in probability as the following theorem
shows
Theorem 10.2.
If X
n
converges to X in L
p
for p > 0, then X
n
converges to X in probability.
Proof
Apply Chebyshevs inequality to f(x) = x
p
, which yields for any > 0,
P([X
n
[ )
p
E([X
n
[
p
) 0, as n .
Theorem 10.3.
If X
n
X in probability, then there exists a subsequence n
k
such that X
n
k
X a.s.
Theorem 10.4.
If X
n
X a.s., then X
n
X in probability.
10.2. Approximation via Riemann Sums
Following Dellacherie and Meyer we shall establish the equivalence of the two constructions
for the quadratic variation by the following theorem which approximates the stochastic
integral via Riemann sums.
Relations to Sums 33
Theorem 10.2.
Let X be a semimartingale, and H a locally bounded previsible CADLAG process starting
from zero. Then
_
t
0
H
s
dX
s
= lim
n
k=0
H
tk2
n
_
X
t(k+1)2
n X
tk2
n
_
u.c.p.
Proof
Let K
s
= H
s
1
st
, and dene the following sequence of simple process approximations
K
n
s
:=
k=0
H
tk2
n1
(tk2
n
,t(k+1)2
n
]
(s).
Clearly this sequence K
n
s
converges pointwise to K
s
. We can decompose the semimartingale
X as X = X
0
+ A
t
+ M
t
where A
t
is of nite variation and M
t
is a continuous local
martingale, both starting from zero. The result that
_
t
0
K
n
s
dA
s
_
t
0
K
s
dA
s
, u.c.p.
is standard from the LebesgueStieltjes theory. Let T
k
be a reducing sequence for the
continuous local martingale M such that M
T
k
is a bounded martingale. Also since K is
locally bounded we can nd a sequence of stopping times S
k
such that K
S
k
is a bounded
previsible process. It therefore suces to prove for a sequence of stopping times R
k
such
that R
k
, then
(K
n
M)
R
k
s
(K M)
R
k
s
, u.c.p..
By Doobs L
2
inequality, and the Ito isometry we have
E
_
((K
n
M) (K M))
2
4E[(K
n
M) (K M)]
2
, Doob L
2
4K
n
K
2
M
, Ito Isometry
4
_
(K
n
s
K
s
)
2
dM)
s
As [K
n
K[ 0 pointwise, and K is bounded, clearly [K
n
K[ is also bounded uniformly
in n. Hence by the Dominated Convergence Theorem for the LebesgueStieltjes integral
_
(K
n
s
K
s
)
2
dM)
s
0 a.s..
Hence, we may conclude
E
_
[(K
n
M) (K M)]
_
2
0, as n .
So
[(K
n
M) (K M)]
0 in L
2
,
Relations to Sums 34
as n ; but this implies that
[(K
n
M) (K M)]
0 in probability.
Hence
[(K
n
M) (K M)] 0 u.c.p.
as required, and putting the two parts together yields
_
t
0
K
n
s
dX
s
_
t
0
K
s
dX
s
, u.c.p.
which is the required result.
This result can now be applied to the construction of the quadratic variation process,
as illustrated by the next theorem.
Theorem 10.3.
The quadratic variation process X)
t
is equal to the following limit in probability
X)
t
= lim
n
k=0
_
X
t(k+1)2
n X
tk2
n
_
2
in probability.
Proof
In the theorem (7.2) establishing the existence of the quadratic variation process, we noted
in () that
A
n
t
= M
2
t
2(H
n
M)
t
.
Now from application of the previous theorem
2
_
t
0
X
s
dX
s
= lim
n
k=0
X
tk2
n
_
X
t(k+1)2
n X
tk2
n
_
.
In addition,
X
2
t
X
2
0
=
k=0
_
X
2
t(k+1)2
n X
2
tk2
n
_
.
The dierence of these two equations yields
A
t
= X
2
0
+ lim
n
k=0
_
X
t(k+1)2
n X
tk2
n
_
2
,
where the limit is taken in probability. Hence the process A is increasing and positive on
the rational numbers, and hence on the whole of R by right continuity.
Remark
The theorem can be strengthened still further by a result of DoleansDade to the eect
that for X a continuous semimartingale
X)
t
= lim
n
k=0
_
X
t(k+1)2
n X
tk2
n
_
2
,
where the limit is in the strong sense in L
1
. This result is harder to prove (essentially the
uniform integrability of the sums must be proven) and this is not done here.
11. It os Formula
Itos Formula is the analog of integration by parts in the stochastic calculus. It is also
the rst place where we see a major dierence creep into the theory, and realise that our
formalism has found a new subtlety in the subject.
More importantly, it is the fundamental weapon used to evaluate Ito integrals; we
shall see some examples of this shortly.
The Ito isometry provides a cleancut denition of the stochastic integral; however it
was originally dened via the following theorem of Kunita and Watanabe.
Theorem (KunitaWatanabe Identity) 11.1.
Let M /
2
and H and K are locally bounded previsible processes. Then (H M)
is
the unique element of L
2
(T
] = E[(H M, N))
]
Moreover we have
(H M), N) = H M, N).
Proof
Consider an elementary function H, so H = Z1
(S,T]
, where Z is an T
S
measurable bounded
random variable, and S and T are stopping times such that S T. It is clear that
E[(H M)
] =E[Z (M
T
M
S
) N
]
=E[Z(M
T
N
T
M
S
N
S
)]
=E[M
(H N)
]
Now by linearity this can be extended to establish the result for all simple functions (in
L). We nally extend to general locally bounded previsible H, by considering a sequence
(provided it exists) of simple functions H
n
such that H
n
H in L
2
(M). Then there exists
a subsequence n
k
such that H
n
k
converges to H is L
2
(N). Then
E
_
(H
n
k
M)
_
E
_
(H M)
_
=E
_
_
(H
n
k
H) M
_
N
_
E
_
[(H
n
k
H) M)]
2
_
_
E(N
2
_
E
_
[(H
n
k
M) (H M)]
2
_
_
E(N
2
)
By construction H
n
k
H in L
2
(M) which means that
H
n
k
H
M
0, as k .
By the Ito isometry
E
_
((H
n
k
H) M)
2
_
= H
n
k
H
2
M
0, as k ,
[35]
Itos Formula 36
that is (H
n
k
M)
(H M)
in L
2
. Hence as N is an L
2
bounded martingale, the right
hand side of the above expression tends to zero as k . Similarly as (H
n
k
N)
(H N)
in L
2
, we see also that
E((H
n
k
N)
) 0, as k .
Hence we can pass to the limit to obtain the result for H.
To prove the second part of the theorem, we shall rst show that
(H N), (K M)) +(K N), (H M)) = 2HKM, N).
By polarisation
M, N) =
M +N) M N)
4
,
also
HK =
(H +K)
2
(H K)
2
4
.
Hence
2(HK M, N)) =
1
8
_
_
(H +K)
2
(H K)
2
M+N) MN)
_
.
Now we use the result that (H M)) = (H
2
M)) which has been proved previously in
theorem (7.9(iii)), to see that
2(HK M, N)) =
1
8
_
(H +K) (M +N)) (H +K) (M N))
(H K) (M +N)) +(H K) (M N))
_
.
Considering the rst two terms
(H +K)(M +N)) (H +K) (M N)) =
=(H +K) M + (H +K) N) (H +K) M (H +K) N)
=4(H +K) M, (H +K) N) by polarisation
=4 (H M, H N) +H M, K N) +K M, H N) +K M, K N)) .
Similarly for the second two terms
(H K)(M +N)) (H K) (M N)) =
=(H K) M + (H K) N) (H K) M (H K) N)
=4(H K) M, (H K) N) by polarisation
=4 (H M, H N) H M, K N) K M, H N) +K M, K N)) .
Itos Formula 37
Adding these two together yields
2(HK M, N)) = (H N), (K M)) +(K N), (H M))
Putting K 1 yields
2H M, N) = H M, N) +M, H N).
So it suces to prove that (H M), N) = M, (H N)), which is equivalent to showing
that
(H M)N (H N)M
is a local martingale (from the denition of covariation process). By localisation it suces
to consider M and N bounded martingales, whence we must check that for all stopping
times T,
E((H M)
T
N
T
) = E((H N)
T
M
T
) ,
but by the rst part of the theorem
E((H M)
) = E((H N)
) ,
which is sucient to establish the result, since
(H M)
T
N
T
=(H M)
T
N
T
(H N)
T
M
T
=(H N)
T
M
T
Corollary 11.2.
Let N, M be continuous local martingales and H and K locally bounded previsible pro
cesses, then
(H N), (K M)) = (HK N, M)).
Proof
Note that the covariation is symmetric, hence
(H N), (K M)) =(H X, (K M)))
=(H (K M), X)))
=(HK M, N)).
We can prove a stochastic calculus analogue of the usual integration by parts formula.
However note that there is an extra term on the right hand side, the covariation of the
processes X and Y . This is the rst major dierence we have seen between the Stochastic
Integral and the usual Lebesgue Integral.
Before we can prove the general theorem, we need a lemma.
Itos Formula 38
Lemma (Parts for Finite Variation Process and a Martingale) 11.3.
Let M be a bounded continuous martingale starting from zero, and V a bounded variation
process starting from zero. Then
M
t
V
t
=
_
t
0
M
s
dV
s
+
_
t
0
V
s
dM
s
.
Proof
For n xed, we can write
M
t
V
t
=
k1
M
k2
n
t
_
V
k2
n
t
V
(k1)2
n
t
_
+
k1
V
(k1)2
n
t
_
M
k2
n
t
M
(k1)2
n
t
_
=
k1
M
k2
n
t
_
V
k2
n
t
V
(k1)2
n
t
_
+
_
t
0
H
n
s
dM
s
,
where H
n
is the previsible simple process
H
n
s
=
k1
V
k2
n
t
1
((k1)2
n
t,k2
n
t]
(s).
These H
n
are bounded and converge to V by the continuity of V , so as n the second
term tends to
_
t
0
V
s
dM
s
,
and by the Dominated Convergence Theorem for LebesgueStieltjes integrals, the second
term converges to
_
t
0
M
s
dV
s
,
as n .
Theorem (Integration by Parts) 11.4.
For X and Y continuous semimartingales, then the following holds
X
t
Y
t
X
0
Y
0
=
_
t
0
X
s
dY
s
+
_
t
0
Y
s
dX
s
+X, Y )
t
.
Proof
It is trivial to see that it suces to prove the result for processes starting from zero.
Hence let X
t
= M
t
+ A
t
and Y
t
= N
t
+ B
t
in DoobMeyer decomposition, so N
t
and M
t
are continuous local martingales and A
t
and B
t
are nite variation processes, all starting
from zero. By localisation we can consider the local martingales M and N to be bounded
martingales and the FV processes A and B to have bounded variation. Hence by the usual
(nite variation) theory
A
t
B
t
=
_
t
0
A
s
dB
s
+
_
t
0
B
s
dA
s
.
Itos Formula 39
It only remains to prove for bounded martingales N and M starting from zero that
M
t
N
t
=
_
t
0
M
s
dN
s
+
_
t
0
N
s
dM
s
+M, N)
t
.
This follows by application of polarisation to corollary (7.3) to the quadratic variation
existence theorem. Hence
(M
t
+A
t
)(N
t
+B
t
) =M
t
N
t
+M
t
B
t
+N
t
A
t
+A
t
B
t
=
_
t
0
M
s
dN
s
+
_
t
0
N
s
dM
s
+M, N)
t
+
_
t
0
M
s
dB
s
+
_
t
0
B
s
dM
s
+
_
t
0
N
s
dA
s
+
_
t
0
A
s
dN
s
+
_
t
0
A
s
dB
s
+
_
t
0
B
s
dA
s
=
_
t
0
(M
s
+A
s
)d(N
s
+B
s
) +
_
t
0
(N
s
+B
s
)d(M
s
+A
s
) +M, N).
Reect for a moment that this theorem is implying another useful closure property of
continuous semimartingales. It implies that the product of two continuous semimartingales
X
t
Y
t
is a continuous semimartingale, since it can be written as a stochastic integrals with
respect to continuous semimartingales and so it itself a continuous semimartingale.
Theorem (Itos Formula) 11.5.
Let f : R
n
R
n
be a twice continuously dierentiable function, and also let X =
(X
1
,X
2
, . . . , X
n
) be a continuous semimartingale in R
n
. Then
f(X
t
) f(X
0
) =
n
i=1
_
t
0
f
x
i
(X
s
)dX
i
s
+
1
2
n
i,j=1
_
t
0
f
x
i
x
j
(X
s
)dX
i
, X
j
)
s
.
Proof
To prove It os formula; rst consider the n = 1 case to simplify the notation. Then let
/ be the collection of C
2
(twice dierentiable) functions f : R R for which it holds.
Clearly / is a vector space; in fact we shall show that it is also an algebra. To do this we
must check that if f and g are in /, then their product fg is also in /. Let F
t
= f(X
t
)
and G
t
= g(X
t
) be the associated semimartingales. From the integration by parts formula
F
t
G
t
F
0
G
0
=
_
t
0
F
s
dG
s
+
_
t
0
G
s
dF
s
+F
s
, G
s
).
However since by assumption f and g are in /, Itos formula may be applied to them
individually, so
_
t
0
F
s
dG
s
=
_
t
0
f(X
s
)
f
x
(X
s
)dX
s
.
Itos Formula 40
Also by the KunitaWatanabe formula extended to continuous local martingales we have
F, G)
t
=
_
t
0
f
(X
s
)g
(X
s
)dX, X)
s
.
Thus from the integration by parts,
F
t
G
t
F
0
G
0
=
_
t
0
F
s
dG
s
+
_
t
0
G
s
dF
s
+
_
t
0
f
(X
s
)g
(X
s
)dX, X)
s
,
=
_
t
0
(F
s
g
(X
s
) +f
(X
s
)G
s
) dX
s
+
1
2
_
t
0
(F
s
g
(X
s
) + 2f
(X
s
) +f
(X
s
)G
s
) dM)
s
.
So this is just what Itos formula states for fg and so Itos formula also applies to fg;
hence fg /.
Since trivially f(x) = x is in /, then as / is an algebra, and a vector space this
implies that / contains all polynomials. So to complete the proof, we must approximate f
by polynomials (which we can do by standard functional analysis), and check that in the
limit we obtain Itos formula.
Introduce a sequence U
n
:= inft : [X
t
[ + X)
t
> n. Hence U
n
is a sequence of
stopping times tending to innity. Now we shall prove Itos formula for twice continuously
dierentiable f restricted to the interval [0, U
n
], so we can consider X as a bounded martin
gale. Consider a polynomial sequence f
k
approximating f, in the sense that for r = 0, 1, 2,
f
(r)
k
f
(r)
uniformly on a compact interval. We have proved that Itos formula holds for
all polynomial, so it holds for f
k
and hence
f
k
(X
tU
n
) f
k
(X
0
) =
_
tU
n
0
f
(X
s
)dX
s
+
1
2
_
tU
n
0
f
k
(X
s
)dX)
s
.
Let the continuous semimartingale X have DoobMeyer decomposition
X
t
= X
0
+M
t
+A
t
,
where M is a continuous local martingale and A is a nite variation process. We can rewrite
the above as
f
k
(X
tU
n
) f
k
(X
0
) =
_
tU
n
0
f
(X
s
)dM
s
+
_
tU
n
0
f
(X
s
)dA
s
+
1
2
_
tU
n
0
f
k
(X
s
)dM)
s
.
since X) = M). On (0, U
n
] the process [X[ is uniformly bounded by n, so for r = 0, 1, 2
from the convergence (which is uniform on the compact interval [0, U
n
]) we obtain
sup
xn
f
(r)
k
f
(r)
0 as k
And from the Ito isometry we get the required convergence.
Itos Formula 41
11.1. Applications of Itos Formula
Let B
t
be a standard Brownian motion; the aim of this example is to establish that:
_
t
0
B
s
dB
s
=
1
2
B
2
t
1
2
t.
This example gives a nice simple demonstration that all our hard work has achieved some
thing. The result isnt the same as that which would be given by the logical extension of
the usual integration rules.
To prove this we apply Itos formula to the function f(x) = x
2
. We obtain
f(B
t
) f(B
0
) =
_
t
0
f
x
(B
s
)dB
s
+
1
2
_
t
0
2
f
x
2
(B
s
)dB, B)
s
,
noting that B
0
= 0 for a standard Brownian Motion we see that
B
2
t
= 2
_
t
0
B
s
dB
s
+
1
2
2ds,
whence we derive that
_
t
0
B
s
dB
s
=
B
2
t
2
t
2
.
For those who have read the foregoing material carefully, there are grounds to complain that
there are simpler ways to establish this result, notably by consideration of the denition of
the quadratic variation process. However the point of this example was to show how Itos
formula can help in the actual evaluation of stochastic integrals; not to establish a totally
new result.
11.2. Exponential Martingales
Exponential martingales play an important part in the theory. Suppose X is a continuous
semimartingale starting from zero. Dene:
Z
t
= exp
_
X
t
1
2
X)
t
_
.
This Z
t
is called the exponential semimartingale associated with X
t
, and it is the
solution of the stochastic dierential equation
dZ
t
= Z
t
dX
t
,
that is
Z
t
= 1 +
_
t
0
Z
s
dX
s
,
so clearly if X is a continuous local martingale, i.e. X /
c
loc
then this implies, by the
stability property of stochastic integration, that Z /
c
loc
also.
Itos Formula 42
Proof
For existence, apply Itos formula to f(x) = exp(x) to obtain
d (exp(Y
t
)) = exp(Y
t
)dY
t
+
1
2
exp(Y
t
)dY, Y )
t
.
Hence
d
_
exp(X
t
1
2
X)
t
)
_
=exp(X
t
1
2
X)
t
)d
_
X
t
1
2
X)
t
_
+
1
2
exp
_
X
t
1
2
X)
t
_
d
_
X
t
1
2
X)
t
, X
t
1
2
X)
t
_
=exp(X
t
1
2
X)
t
)dX
t
1
2
exp
_
X
t
1
2
X)
t
_
dX)
t
+
1
2
exp
_
X
t
1
2
X)
t
_
dX)
t
=Z
t
dX
t
Hence Z
t
certainly solves the equation. Now to check uniqueness, dene
Y
t
= exp
_
X
t
+
1
2
X)
t
_
,
we wish to show that for every solution of the Stochastic Dierential Equation Z
t
Y
t
is a
constant. By a similar application of Itos formula
dY
t
= Y
t
dX
t
+Y
t
dX)
t
,
whence by integration by parts (alternatively consider Ito applied to f(x, y) = xy),
d(Z
t
Y
t
) =Z
t
dY
t
+Y
t
dZ
t
+Z, Y )
t
,
=Z
t
(Y
t
dX
t
+Y
t
dX)
t
) +Y
t
Z
t
dX
t
+ (Y
t
Z
t
)dX)
t
,
=0.
So Z
t
Y
t
is a constant, hence the unique solution of the stochastic dierential equation
dZ
t
= Z
t
dX
t
, with Z
0
= 1, is
Z
t
= exp
_
X
t
1
2
X)
t
_
.
Itos Formula 43
Example
Let X
t
= B
t
, for an arbitrary scalar . Clearly X
t
is a continuous local martingale, so
the associated exponential martingale is
M
t
= exp
_
B
t
1
2
2
t
_
.
It is clear that the exponential semimartingale of a real valued martingale must be
nonnegative, and thus by application of Fatous lemma we can show that it is a super
martingale, thus E(M
t
) 1 for all t.
Theorem 11.6.
Let M be a nonnegative local martingale, such that EM
0
< then M is a supermartin
gale.
Proof
Let T
n
be a reducing sequence for M
n
M
0
, then for t > s 0,
E(M
tT
n
[T
s
) = E(M
0
[T
s
) +E(M
tT
n
M
0
[T
s
)
= M
0
+M
sT
n
M
0
= M
sT
n
.
Now by application of the conditional form of Fatous lemma
E(M
t
[F
s
) = E(liminf
n
M
tT
n
[T
s
) liminf
n
E(M
tT
n
[T
s
) = M
s
.
Thus M is a supermartingale as required.
Checking for a Martingale
The general exponential semimartingale is useful, but in many applications, not the least
of which will be Girsanovs formula an actual Martingale will be needed. How do we go
about checking if a local martingale is a martingale anyway? It will turn out that there are
various methods, some of which crop up in the section on ltration. First I shall present a
simple example and then prove a more general theorem. A common error is to think that
it is sucient to show that a local martingale is locally bounded in L
2
to show that it is
a martingale this is not sucient as should be made clear by this example!
The exponential of B
t
We continue our example from above. Let M
t
= exp(B
t
1/2
2
t) be the exponential
semimartingale associated with a standard Brownian Motion B
t
, starting from zero. By
the previous argument we know that
M
t
= 1 +
_
t
0
M
s
dB
s
,
hence M is a local martingale. Fix T a constant time, which is of course a stopping time,
then B
T
is an L
2
bounded martingale (E(B
T
t
)
2
= t T T). We then show that M
T
is
Itos Formula 44
in L
2
(B
T
) as follows
M
T

B
= E
_
_
T
0
M
2
s
dB)
s
_
,
= E
_
_
T
0
M
2
s
ds
_
,
E
_
_
T
0
exp(2B
s
)ds
_
,
=
_
T
0
E(exp(2B
s
)) ds,
=
_
T
0
exp(2
2
s
2
)ds < .
In the nal equality we use that fact that B
s
is distributed as N(0, s), and we use the char
acteristic function of the normal distribution. Thus by the integration isometry theorem
we have that (M
T
B)
t
is an L
2
bounded martingale. Thus for every such T, Z
T
is an L
2
bounded martingale, which implies that M is a martingale.
The Exponential Martingale Inequality
We have seen a specic proof that a certain (important) exponential martingale is a true
martingale, we now show a more general argument.
Theorem 11.7.
Let M be a continuous local martingale, starting from zero. Suppose for each t, there exists
a constant K
t
such that M)
t
< a.s., then for every t, and every y > 0,
P
_
sup
0st
M
s
> y
_
exp(y
2
/2K
t
).
Furthermore, the associated exponential semimartingale Z
t
= exp(M
t
1/2
2
M)
t
) is a
true martingale.
Proof
We have already noted that the exponential semimartingale Z
t
is a supermartingale, so
E(Z
t
) 1 for all t 0, and hence for > 0 and y > 0,
P
_
sup
st
M
s
> y
_
P
_
sup
st
Z
s
> exp(y 1/2
2
K
t
)
_
,
exp(y + 1/2
2
K
t
).
Optimizing over now gives the desired result. For the second part, we establish the
Itos Formula 45
following bound
E
_
sup
0st
Z
s
_
E
_
exp
_
sup
0st
Z
s
__
,
_
0
P[sup 0 s tZ
s
log ] d,
1 +
_
1
exp
_
(log )
2
/2K
t
_
d < .
()
We have previously noted that Z is a local martingale; let T
n
be a reducing sequence for
Z, hence Z
T
n
is a martingale, hence
E[Z
tT
n
[T
s
] = Z
sT
n
. ()
We note that Z
t
is dominated by exp( sup
0st
Z
s
), and thus by our bound we can
apply the dominated convergence theorem to () as n to establish that Z is a true
martingale.
Corollary 11.8.
For all , > 0,
P
_
sup
t0
M
t
& M)
_
exp(
2
/2).
Proof
Set T = inft 0 : M
t
, the conditions of the previous theorem now apply to M
T
,
with K
t
= .
From this corollary, it is clear that if H is any bounded previsible process, then
exp
__
t
0
H
s
dB
s
1
2
_
t
0
[H
s
[
2
ds
_
is a true martingale, since this is the exponential semimartingale associated with the process
_
HdB.
Corollary 11.9.
If the bounds K
t
on M) are uniform, that is if K
t
C for all t, then the exponential
martingale is Uniformly Integrable. We shall use the useful result
_
0
P(X log )d =
_
0
E
_
1
e
X
_
d = E
_
0
1
e
X
d = E(e
X
).
Proof
Note that the bound () extends to a uniform bound
E
_
sup
t0
Z
t
_
1 +
_
1
exp
_
(log )
2
/2C
_
d < .
Hence Z is bounded in L
t0
be continuous local martingales starting from zero for i = 1, . . . , n. Then
B
t
= (B
1
t
, . . . , B
n
t
) is a Brownian motion with respect to (, T, P) adapted to the ltration
T
t
, if and only i
B
i
, B
j
)
t
=
ij
t i, j 1, . . . , n.
Proof
In these circumstances it follows that the statement B
t
is a Brownian Motion is by denition
equivalent to stating that B
t
B
s
is independent of T
s
and is distributed normally with
mean zero and covariance matrix (t s)I.
Clearly if B
t
is a Brownian motion then the covariation result follows trivially from the
denitions. Now to establish the converse, we assume B
i
, B
j
)
t
=
ij
t for i, j 1, . . . , n,
and shall prove B
t
is a Brownian Motion.
Observe that for xed R
n
we can dene M
t
by
M
t
:= f(B
t
, t) = exp
_
i(, x) +
1
2
[[
2
t
_
.
By application of Itos formula to f we obtain (in dierential form using the Einstein
summation convention)
d (f(B
t
, t)) =
f
x
j
(B
t
, t)dB
j
t
+
f
t
(B
t
, t)dt +
1
2
2
f
x
j
x
k
(B
t
, t)dB
j
, B
k
)
t
,
=i
j
f(B
t
, t)dB
j
t
+
1
2
[[
2
f(B
t
, t)dt
1
2
jk
f(B
t
, t)dt
=i
j
f(B
t
, t)dB
j
t
.
Hence
M
t
= 1 +
_
t
0
d(f(B
t
, t)),
and is a sum of stochastic integrals with respect to continuous local martingales and is
hence itself a continuous local martingale. But note that for each t,
[M
t
[ =
_
e
1
2

2
t
_
<
Hence for any xed time t
0
, (M
t
0
)
t
satises
[(M
t
0
)
t
[ [(M
t
0
)
[ < ,
[46]
Levy Characterisation of Brownian Motion 47
and so is a bounded local martingale; hence (M
t
0
)
t
) is a martingale. Hence M
t
0
is a genuine
martingale. Thus for 0 s < t we have
E(exp (i(, B
t
B
s
)) [T
s
) = exp
_
1
2
(t s) [[
2
_
a.s.
However this is just the characteristic function of a normal random variable following
N(0, t s); so by the Levy character theorem B
t
B
s
is a N(0, t s) random variable.
13. Time Change of Brownian Motion
This result is one of frequent application, essentially it tells us that any continuous local
martingale starting from zero, can be written as a time change of Brownian motion. So
modulo a time change a Brownian motion is the most general kind of continuous local
martingale.
Proposition 13.1.
Let M be a continuous local martingale starting from zero, such that M)
t
as t .
Then dene
s
:= inft > 0 : M)
t
> s.
Then dene
A
s
:= M
s
.
(i) This
s
is an T stopping time.
(ii) M)
s
= s.
(iii) The local martingale M can be written as a time change of Brownian Motion as
M
t
= B
M)
t
. Moreover the process
A
s
is an
T
s
adapted Brownian Motion, where
T
s
is
the timechanged algebra i.e.
T
s
= T
s
.
Proof
We may assume that the map t M)
t
is strictly increasing. Note that the map s
s
is
the inverse to t M)
t
. Hence the results (i),(ii) and (iii).
Dene
T
n
:= inft : [M[
t
> n,
[U
n
:= M)
T
n
.
Note that from these denitions
tU
n
=infs > 0 : M)
s
> t U
n
=infs > 0 : M)
s
> t M)
T
n
=T
n
t
So
A
U
n
s
=
A
sU
n
= M
T
n
t
.
Now note that U
n
is an
T
t
stopping time, since consider
U
n
t M)
T
n
t T
n
t
,
the latter event is clearly T
t
measurable i.e. it is
F
t
measurable, so U
n
is a
T
t
stopping
time. We may now apply the optional stopping theorem to the UI martingale M
T
n
, which
yields
E
_
A
U
n
t
[T
s
_
=E
_
A
tU
n
[
T
s
_
= E
_
M
T
n
t
[
T
s
_
=E
_
M
T
n
t
[T
s
_
= M
T
n
s
=
A
U
n
s
.
[48]
Time Change of Brownian Motion 49
So
A
t
is a
T
t
local martingale. But we also know that (M
2
M))
T
n
is a UI martingale,
since M
T
n
is a UI martingale. By the optional stopping theorem, for 0 < r < s we have
E
_
A
2
sU
n
(s U
n
)[
T
r
_
= E
__
_
M
T
n
s
_
2
M)
s
T
n
_
[T
r
_
=E
_
_
M
2
s
M)
s
_
T
n
[T
r
_
=
_
M
2
r
M)
r
_
T
n
=
A
2
rU
n
(r U
n
).
Hence
A
2
t is a
T
t
local martingale. Before we can apply Levys characterisation theorem
we must check that
A is continuous; that is we must check that for almost every that
M is constant on each interval of constancy of M). By localisation it suces to consider
M a square integrable martingale, now let q be a positive rational, and dene
S
q
:= inft > q : M)
t
> M)
q
,
then it is enough to show that M is constant on [q, S
q
). But M
2
M) is a martingale,
hence
E
_
_
M
2
S
q
M)
S
q
_
2
[T
q
_
=M
2
q
M)
q
=M
2
q
M)
S
q
, as M)
q
= M)
S
q
.
Hence
E
_
_
M
S
q
M
q
_
2
[T
q
_
= 0,
which establishes that
A is continuous.
Thus
A is a continuous
T
t
adapted martingale with
A
s
)
s
= s and so by the Levy
characterisation theorem
A
s
is a Brownian Motion.
13.1. Gaussian Martingales
The time change of Brownian Motion can be used to prove the following useful theorem.
Theorem 13.2.
If M is a continuous local martingale starting from zero, and M)
t
is deterministic, that
is if we can nd a deterministic function f taking values in the nonnegative real numbers
such that M)
t
= f(t) a.e., then M is a Gaussian Martingale (i.e. M
t
has a Gaussian
distribution for almost all t).
Proof
Note that by the time change of Brownian Motion theorem, we can write M
t
as a time
change of Brownian Motion through
M
t
= B
M)
t
,
where B is a standard Brownian Motion. By hypothesis M)
t
= f(t), a deterministic
function for almost all t, hence for almost all t,
M
t
= B
f(t)
,
Time Change of Brownian Motion 50
but the right hand side is a Gaussian random variable following N(0, f(t)). Hence M is a
Gaussian Martingale, and at time t it has distribution given by N(0, M)
t
).
As a corollary consider the stochastic integral of a purely deterministic function with
respect to a Brownian motion.
Corollary 13.3.
Let g(t) be a deterministic function of t, then M dened by
M
t
:=
_
t
0
f(s)dB
s
,
satises
M
t
N
_
0,
_
t
0
[f(s)[
2
ds
_
.
Proof
From the denition of M via a stochastic integral with respect to a continuous martingale,
it is clear that M is a continuous local martingale, and by the KunitaWatanabe result,
the quadratic variation of M is given by
M)
t
=
_
t
0
[f(s)[ds,
hence the result follows.
This result can also be established directly in a fashion which is very similar to the
proof of the Levy characterisation theorem. Consider Z dened via
Z
t
= exp
_
iM
t
+
1
2
2
M)
t
_
,
as in the Levy characterisation proof, we see that this is a continuous local martingale,
and by boundedness furthermore is a martingale, and hence
E(Z
0
) = E(Z
t
),
whence
E(exp(iM
t
)) = E
_
exp
_
1
2
2
_
t
0
f(s)
2
ds
__
which is the characteristic function of the appropriate normal distribution.
14. Girsanovs Theorem
Girsanovs theorem is an element of stochastic calculus which does not have an analogue
in standard calculus.
14.1. Change of measure
When we wish to compare two measures P and Q, we dont want either of them simply to
throw information away; since when they are positive they can be related by the Radon
Nikodym derivative; this motivates the following denition of equivalence of two measures.
Denition 14.1.
Two measures P and Q are said to be equivalent if they operate on the same sample space,
and if A is any event in the sample space then
P(A) > 0 Q(A) > 0.
In other words P is absolutely continuous with respect to Q and Q is absolutely continuous
with respect to P.
Theorem 14.2.
If P and Q are equivalent measures, and X
t
is an T
t
adapted process then the following
results hold
E
Q
(X
t
) = E
P
_
dQ
dP
X
t
_
,
E
Q
(X
t
[T
s
) = L
1
s
E
P
(L
t
X
t
[T
s
) ,
where
L
s
= E
P
_
dQ
dP
T
s
_
.
Here L
t
is the RadonNikodym derivative of Q with respect to P. The rst result basically
shows that this is a martingale, and the second is a continuous time version of Bayes
theorem.
Proof
The rst part is basically the statement that the RadonNikodym derivative is a martingale.
This follows because the measures P and Q are equivalent, but this will not be proved in
detail here. Let Y be an T
t
measurable random variable, such that E
Q
([Y [) < . We shall
prove that
E
Q
(Y [T
s
) =
1
L
s
E
P
[Y L
t
[T
s
] a.s. (P and Q).
Then for any A T
s
, using the denition of conditional expectation we have that
E
Q
_
1
A
1
L
s
E
P
[Y L
t
[T
s
]
_
=E
P
(1
A
E
P
[Y L
t
[T
s
])
=E
P
[1
A
Y L
t
] = E
Q
[1
A
Y ] .
Substituting Y = X
t
gives the desired result.
[51]
Girsanovs Theorem 52
Theorem (Girsanov).
Let M be a continuous local martingale, and let Z be the associated exponential martingale
Z
t
= exp
_
M
t
1
2
M)
t
_
.
If Z is uniformly integrable, then a new measure Q, equivalent to P may be dened by
dQ
dP
= Z
.
Then if X is a continuous P local martingale, X X, M) is a Q local martingale.
Proof
Since Z
[T
t
). Hence Q constructed thus is
a probability measure which is equivalent to P. Now consider X, a P local martingale.
Dene a sequence of stopping times which tend to innity via
T
n
:= inft 0 : [X
t
[ n, or [X, M)
t
[ n.
Now consider the process Y dened via
Y := X
T
n
X
T
n
, M).
By Itos formula for 0 t T
n
, remembering that dZ
t
= Z
t
dM
t
as Z is the exponential
martingale associated with M,
d(Z
t
Y
t
) =Z
t
dY
t
+Y
t
dZ
t
+ dZ, Y )
t
=Z
t
(dX
t
dX, M)
t
) +Y
t
Z
t
dM
t
+ dZ, Y )
t
=Z
t
(dX
t
dX, M)
t
) + (X
t
X, M)
t
)Z
t
dM
t
+Z
t
dX, M)
t
=(X
t
X, M)
t
)Z
t
dM
t
+Z
t
dX
t
Where the result dZ, Y )
t
= Z
t
dX, M)
t
follows from the KunitaWatanabe theorem.
Hence ZY is a Plocal martingale. But since Z is uniformly integrable, and Y is bounded
(by construction of the stopping time T
n
), hence ZY is a genuine Pmartingale. Hence for
s < t and A T
s
, we have
E
Q
[(Y
t
Y
s
)1
A
] = E[Z
(Y
t
Y
s
)1
A
] = E[(Z
t
Y
t
Z
s
Y
s
)1
A
] = 0,
hence Y is a Q martingale. Thus XX, M) is a Q local martingale, since T
n
is a reducing
sequence such that (X X, M))
T
n
is a Qmartingale, and T
n
as n .
Corollary 14.3.
Let W
t
be a P Brownian motion, then
W
t
:= W
t
W, M)
t
is a Q Brownian motion.
Proof
Use Levys characterisation of Brownian motion to see that since
W
t
is continuous and
W,
W)
t
= W, W)
t
= t, since W
t
is a P Brownian motion, then
W is a Q Brownian
motion.
15. Brownian Martingale Representation Theorem
The following theorem has many applications, for example in the rigorous study of math
ematical nance, even though the result is purely an existence theorem. The Malliavin
calculus oers methods by which the process H in the following theorem can be stated
explicitly, but these methods are beyond the scope of these notes!
Theorem 15.1.
Let B
t
be a Brownian Motion on R
n
and (
t
is the usual augmentation of the ltration
generated by B
t
. If Y is L
2
integrable and is measurable with respect to (
then there
exists a previsible (
t
measurable process H
s
uniquely dened up to evanescence such that
E(Y [(
t
) = E(Y ) +
_
t
0
H
s
dB
s
(1)
The proof of this result can seem hard if you are not familiar with functional analysis
style arguments. The outline of the proof is to describe all Y s which are (
T
measurable
which cannot be represented in the form (1) as belonging to the orthogonal complement of
a space. Then we show that for Z in this orthogonal complement that E(ZX) = 0 for all X
in a large space of (
T
measurable functions. Finally we show that this space is suciently
big that actually we have proved this for all (
T
measurable functions, which includes Z so
E(Z
2
) = 0 and hence Z = 0 a.s. and we are done!
Proof
Without loss of generality prove the result in the case EY = 0 where Y is L
2
integrable
and measurable with respect to (
T
for some constant T > 0.
Dene the space
L
2
T
(B) =
_
H : H is (
t
previsible and E
_
_
T
0
H
s

2
ds
_
<
_
Consider the stochastic integral map
I : L
2
T
(B) L
2
((
T
)
dened by
I(H) =
_
T
0
H
s
dB
s
.
As a consequence of the Ito isometry theorem, this map is an isometry. Hence the image V
under I of the Hilbert space L
2
T
(B) is complete and hence a closed subspace of L
2
0
((
T
) =
H L
2
((
T
) : EH = 0. The theorem will be proved if we can establish that the image is
the whole space.
We follow the usual approach in such proofs; consider the orthogonal complement of
V in L
2
0
((
T
) and we aim to show that every element of this orthogonal complement is zero.
Suppose that Z is in the orthogonal complement of L
2
0
((
T
), thus
E(ZX) = 0 for all X L
2
0
((
T
) (2)
[53]
Brownian Martingale Representation Theorem 54
We can dene Z
t
= E(Z[(
t
) which is an L
2
bounded martingale. We know that the sigma
eld (
0
is trivial by the 01 law therefore
Z
0
= E(Z[(
0
) = EZ = 0.
Let H L
2
(B) let N
T
= I(H); we may dene N
t
= E(N
T
[(
t
) for 0 t T. Clearly
N
T
V as it is the image under I of some H.
Let S be a stopping time such that S T then
N
S
= E(N
T
[(
S
) = E
_
_
S
0
H
s
dB
s
+
_
T
S
H
s
dB
s
(
S
_
= I(H1
(0,S]
),
so consequently N
S
V . The orthogonality relation (2) then implies that E(ZN
S
) = 0.
Thus using the martingale property of Z,
E(ZN
S
) = E(E(ZN
S
[(
S
)) = E(N
S
E(Z[(
S
)) = E(Z
S
N
S
) = 0
Since Z
T
and N
T
are square integrable, it follows that Z
t
N
t
is a uniformly integrable
martingale.
Since the stochastic exponential of a process may be written as
M
t
= c(i B
t
) = exp
_
i B
t
+
1
2
[ [
2
t
_
=
_
t
0
iM
t
dB
t
,
such a process can be taken as H = i M
t
in the denition of N
T
and by the foregoing
argument we see that Z
t
M
t
is a martingale. Thus
Z
s
M
s
= E(Z
t
M
t
[(
s
) = E
_
Z
t
exp
_
i B
t
+
1
2
[ [
2
t
_
(
s
_
Thus
Z
s
exp
_
1
2
[ [
2
(t s)
_
= E( Z
t
exp (i (B
t
B
s
)[ (
s
)) .
Consider a partition 0 < t
1
< t
2
< < t
m
T, and by repeating the above argument,
conditioning on each (
t
j
we establish that
E
_
_
Z
T
exp
_
_
i
j
j
(B
t
j
B
t
j1
)
_
_
_
_
= E
_
Z
0
exp
_
1
2
(t
j
t
j1
)[
j
[
2
__
= 0,
(3)
where the last equality follows since Z
0
= 0.
This is true for any choices of
j
R
n
for j = 1, . . . n. The complex valued functions
dened on (R
n
)
m
by
P
(r)
(x
1
, . . . , x
m
) =
K
(n)
k=1
c
(r)
k
exp
_
_
i
m
j=1
a
(r)
j,k
x
j
_
_
Brownian Martingale Representation Theorem 55
clearly separate points (i.e. for given distinct points we can choose coecients such that the
functions have distinct values at these points), form a linear space and are closed under
complex conjugation. Therefore by the StoneWeierstass theorem (see [Bollobas, 1990]),
their uniform closure is the space of complex valued functions (recall that the complex
variable form of this theorem only requires local compactness of the domain).
Therefore we can approximate any continuous bounded complex valued function f :
(R
n
)
m
C by a sequence of such Ps. But we have already shown in (3) that
E
_
Z
T
P
(r)
(B
t
1
, . . . , B
t
n
)
_
= 0
Hence by uniform approximation we can extend this to any f continuous, bounded
E(Z
T
f(B
t
1
, . . . , B
t
n
)) = 0.
Now we use the monotone class framework; consider the class H such that for H H,
E(Z
T
H) = 0
This calH is a vector space, and contains the constant one since E(Z) = 0. The fore
going argument shows that it contains all H measurable with respect to the sigma eld
(B
t
1
, . . . , B
t
n
) with 0 < t
1
< t
2
< < t
n
T. Thus the monotone class theorem
implies that it contains all functions which are measurable with respect to (
T
.
The function Z
T
(
T
, and we have shown E(Z
T
X) = 0 for X (
T
. Thus we can
take X = Z
T
whence E(Z
2
T
) = 0 which implies that Z
T
= 0 a.s.. This establishes the
desired result.
The reader should examine the latter part of the proof carefully; it is in fact related
to the proof that the set
_
exp
_
i
_
t
0
s
dB
s
_
: L
([0, t], R
m
)
_
is total in L
1
. A set S is said to be total if E(af) = 0 for all a S implies a = 0 a.s.. The
full proof of this result will reappear in a more abstract form in the stochastic ltering
section of these notes.
16. Stochastic Dierential Equations
Stochastic dierential equations arise naturally in various engineering problems, where the
eects of random noise perturbations to a system are being considered. For example in
the problem of tracking a satelite, we know that its motion will obey Newtons law to a
very high degree of accuracy, so in theory we can integrate the trajectories from the initial
point. However in practice there are other random eects which perturb the motion.
The variety of SDE to be considered here describes a diusion process and has the
form
dX
t
= b(t, X
t
) +(t, X
t
)dB
t
, ()
where b
i
(x, t), and
ij
(t, x) for 1 i d and 1 j r are Borel measurable functions.
In practice such SDEs generally occur written in the Statonowich form, but as we have
seen the Ito form has numerous calculational advantages (especially the fact that local
martinagles are a closed class under the It o integral), so it is conventional to transform the
SDE to the Ito form before proceeding.
Strong Solutions
A strong solution of the SDE () on the given probability space (, T, P) with initial
condition is a process (X
t
)
t0
which has continuous sample paths such that
(i) X
t
is adapted to the augmented ltration generated by the Brownian motion B and
initial condition , which is denoted T
t
.
(ii) P(X
0
= ) = 1
(iii) For every 0 t < and for each 1 i d and 1 j r, then the folllowing holds
almost surely
_
t
0
_
[b
i
(s, X
s
)[ +
2
ij
(s, X
s
)
_
ds < ,
(iv) Almost surely the following holds
X
t
= X
0
+
_
t
0
b(s, X
s
)ds +
_
t
0
(s, X
s
)dB
s
.
Lipshitz Conditions
Let   denote the usual Euclidean norm on R
d
. Recall that a function f is said to be
Lipshitz if there exists a constant K such that
f(x) f(y) Kx y,
we shall generalise the norm concept to a (d r) matrix by dening

2
=
d
i=1
r
j=1
2
ij
.
The concept of Liphshitz continuity can be extended to that of local Lipshitz continuity,
by requiring that for each n there exists K
n
, such that for all x and y such that x n
and y n then
f(x) f(y) K
n
x y.
[56]
Stochastic Dierential Equations 57
Strong Uniqueness of Solutions
Theorem (Uniqueness) 16.1.
Suppose that b(t, x) and (t, x) are locally Lipshitz continuous in the spatial variable (x).
That is for every n 1 there exists a constant K
n
> 0 such that for every t 0, x n
and y n the following holds
b(t, x) b(t, y) +(t, x) (t, y) K
n
x y.
Then strong uniqueness holds for the pair (b, ), which is to say that if X and
X are two
strong solutions of () relative to B with initial condition then X and
X are indistin
guishable, that is
P
_
X
t
=
X
t
t : 0 t <
_
= 1.
The proof of this result is importanty inasmuch as it illustrates the rst example of a
technique of bounding which recurs again and again throughout the theory of stochastic
dierential equations. Therefore I make no apology for spelling the proof out in excessive
detail, as it is most important to understand exactly where each step comes from!
Proof
Suppose that X and
X are strong solutions of (), relative to the same brownian motion
B and initial condition on the same probability space (, T, P). Dene a sequence of
stopping times
n
= inft 0 : X
t
 n, and
n
= inft 0 : Y
t
 n.
Now set S
n
= min(
n
,
n
). Clearly S
n
is also a stopping time, and S
n
a.s. (P) as
n . These stopping times are only needed because b and are being assumed merely
to be locally Lipshitz. If they are assumed to be Lipshitz, as will be needed in the existence
part of the proof, then this complexity may be ignored.
Hence
X
tS
n
X
tS
n
=
_
tS
n
0
_
b(u, X
u
) b(u,
X
u
)
_
du
+
_
tS
n
0
_
(u, X
u
) (u,
X
u
)
_
dW
u
.
Now we consider evaluating EX
tS
n
X
tS
n

2
, the rst stage follows using the identity
(a +b)
2
4(a
2
+b
2
),
EX
tS
n
X
tS
n

2
4E
_
_
tS
n
0
_
b(u, X
u
) b(u,
X
u
)
_
du
_
2
+ 4E
_
_
tS
n
0
_
(u, X
u
) (u,
X
u
)
_
dW
u
_
2
Stochastic Dierential Equations 58
Considering the second term, we use the Ito isometry which we remember states that
(H M)
2
= H
M
, so
E
_
_
tS
n
0
_
(u, X
u
) (u,
X
u
)
_
dW
u
_
2
=E
_
_
tS
n
0
[(u, X
u
) (u,
X
u
)[
2
du
_
The classical Holder inequality (in the form of the Cauchy Schwartz inequality) for
Lebsgue integrals which states that for p, q (1, ), with p
1
+ q
1
= 1 the following
inequality is satised.
_
[f(x)g(x)[d(x)
__
[f(x)[
p
d(x)
_
1/p
__
[g(x)[
q
d(x)
_
1/q
This result may be applied to the other term, taking p = q = 2 which yields
E
_
_
tS
n
0
_
b(u, X
u
) b(u,
X
u
)
_
du
_
2
E
_
_
tS
n
0
b(u, X
u
) b(u,
X
u
)
du
_
2
E
_
_
tS
n
0
1ds
_
tS
n
0
_
b(u, X
u
) b(u,
X
u
)
_
2
ds
_
E
_
t
_
tS
n
0
_
b(u, X
u
) b(u,
X
u
)
_
2
_
Thus combining these two useful inequalities and using the nth local Lipshitz relations we
have that
EX
tS
n
X
tS
n

2
4tE
_
_
tS
n
0
_
b(u, X
u
) b(u,
X
u
)
_
2
_
+ 4E
_
_
tS
n
0
[(u, X
u
) (u,
X
u
)[
2
du
_
4(T + 1)K
2
n
E
_
t
0
_
X
uS
n
X
uS
n
_
2
du
Now by Gronwalls lemma, which in this case has a zero term outside of the integral, we
see that EX
tS
n
X
tS
n

2
= 0, and hence that P(X
tS
n
=
X
tS
n
) = 1 for all t < .
That is these two processes are modications, and thus indistinguishable. Letting n
we see that the same is true for X
t
t0
and
X
t
t0
.
Now we impose Lipshitz conditions on the functions b and to produce an existence
result. The following form omits some measure theoretic details which are very important;
for a clear treatment see Chung & Williams chapter 10.
Stochastic Dierential Equations 59
Theorem (Existence) 16.2.
If the coecients b and satisfy the global Lipshitz conditions that for all u, t
b(u, x) b(u, y) K[x y[, [(t, x) (t, y)[ K[x y[,
and additionally the bounded growth condition
[b(t, x)[
2
+[(t, x)[
2
K
2
(1 +[x[
2
)
Fix a probability space (, T, P). Let be a random valued vector, independent of the
Brownian Motion B
t
, with nite second moment. Let T
t
be the augmented ltration as
sociated with the Brownian Motion B
t
and . Then there exists a continuous, adapted
process X which is a strong solution of the SDE with initial condition . Additionally this
process is square integrable: for each T > 0 there exists C(K, T) such that
E[X
t
[
2
C
_
1 +E[[
2
_
e
Ct
,
for 0 t T.
Proof
This proof proceeds by Picard iteration through a map F, analogously to the deterministic
case to prove the existence of solutions to rst order ordinary dierential equations. This is
a departure from the more conventional proof of this result. Let F be a map from the space
(
T
of continuous adapted processes X from [0, T] to R, such that E
_
_
sup
tT
X
t
_
2
_
<
. Dene X
(k)
t
recursively, with X
(0)
t
= , and
X
(k+1)
t
= F(X
k
)
t
= +
_
t
0
b(s, X
(k)
s
)ds +
_
t
0
(s, X
(k)
s
)dB
s
[Note: we have left out checking that the image of X under F is indeed adapted!] Now note
that using (a+b)
2
2a
2
+ab
2
, we have using the same bounds as in the uniqueness result
that
E
_
_
sup
0tT
F(X)
t
F(Y )
t
_
2
_
2E
_
_
sup
tT
_
T
0
((X
s
) (Y
s
)) dB
s
2
_
_
+ 2E
_
_
sup
tT
_
T
0
(b(X
s
) b(Y
s
)) ds
2
_
_
2K
2
(4 +T)
_
T
0
E
_
_
sup
tT
[X
t
Y
t
[
2
_
2
_
dt.
By induction we see that for each T we can prove
E
_
_
sup
tT
F
n
(X) F
n
(Y )
_
2
_
C
n
T
n
n!
E
_
_
sup
tT
X
t
Y
t
_
2
_
Stochastic Dierential Equations 60
So by taking n suciently large we have that F
n
is a contraction mapping and so by the
contraction mapping theorem, F
n
mapping (
T
to itself has a xed point, which must be
unique, call it X
(T)
. Clearly from the uniqueness part X
(T)
t
= X
(T
)
t
for t T T
a.s.,
and so we may consistently dene X ( by
X
t
= X
(N)
t
for t N, N N,
which solves the SDE, and has already been shown to be unique.
17. Relations to Second Order PDEs
The aim of this section is to show a rather surprising connection between stochastic dier
ential equations and the solution of second order partial dierential equations. Surprising
though the results may seem they often provide a viable route to calculating the solutions
of explicit PDEs (an example of this is solving the BlackScholes Equation in Option Pric
ing, which is much easier to solve via stochastic methods, than as a second order PDE).
At rst this may well seem to be surprising since one problem is entirely deterministic and
the other in inherently stochastic!
17.1. Innitesimal Generator
Consider the following ddimensional SDE,
dX
t
=b(X
t
)dt +(X
t
)dB
t
,
X
0
=x
0
where is a d d matrix with elements =
ij
. This SDE has innitesimal generator
A, where
A =
d
j=1
b
k
(X
t
)
x
j
+
1
2
d
i=1
d
j=1
d
k=1
ik
(X
t
)
kj
(X
t
)
2
x
i
x
j
.
It is conventional to set
a
ij
=
d
k=1
ik
kj
,
whence A takes the simpler form
A =
d
j=1
b
j
(X
t
)
x
j
+
1
2
d
i=1
d
j=1
a
ij
(X
t
)
2
x
i
x
j
.
Why is the denition useful? Consider application of Itos formula to f(X
t
), which
yields
f(X
t
) f(X
0
) =
_
t
0
d
j=1
f
x
j
(X
s
)dX
s
+
1
2
_
t
0
d
i=1
d
j=1
2
f
x
i
x
j
(X
s
)dX
i
, X
j
)
s
.
Substituting for dX
t
from the SDE we obtain,
f(X
t
) f(X
0
) =
_
t
0
_
_
d
j=1
b
j
(X
s
)
f
x
j
(X
s
) +
1
2
d
i=1
d
j=1
d
k=1
ik
kj
2
f
x
i
x
j
(X
s
)
_
_
dt
+
_
t
0
d
j=1
ij
(X
s
)
f
x
j
(X
s
)dB
s
=
_
t
0
Af(X
s
)ds +
_
t
0
d
j=1
ij
(X
s
)
f
x
j
(X
s
)dB
s
[61]
Relations to Second Order PDEs 62
Denition 17.1.
We say that X
t
satises the martingale problem for A, if X
t
is T
t
adapted and
M
t
= f(X
t
) f(X
0
)
_
t
0
Af(X
s
)ds,
is a martingale for each f C
2
c
(R
d
).
It is simple to verify from the foregoing that any solution of the associated SDE
solves the martingale problem for A. This can be generalised if we consider test functions
C
2
(R
+
R
d
, R), and dene
M
t
:= (t, X
t
) (0, X
0
)
_
t
0
_
s
+A
_
(s, X
s
)ds.
then M
t
is a local martingale, for X
t
a solution of the SDE associated with the innitesimal
generator A. The proof follows by an application of Itos formula to M
t
, similar to that
of the above discussion.
17.2. The Dirichlet Problem
Let be a subspace of R
d
with a smooth boundary . The Dirichlet Problem for f is
dened as the solution of the system
Au + = 0 on ,
u = f on .
Where A is a second order partial dierential operator of the form
A =
d
j=1
b
j
(X
t
)
x
j
+
1
2
d
i=1
d
j=1
a
ij
(X
t
)
2
x
i
x
j
,
which is associated as before to an SDE. This SDE will play an important role in what is
to follow.
A simple example of a Dirichlet Problem is the solution of the Laplace equation in
the disc, with Dirichlet boundary conditions on the boundary, i.e.
2
u = 0 on D,
u = f on D.
Relations to Second Order PDEs 63
Theorem 17.2.
For each f C
2
b
() there exists a unique u C
2
b
() solving the Dirichlet problem for f.
Moreover there exists a continuous function m : (0, ) such that for all f C
2
b
()
this solution is given by
u(x) =
_
m(x, y)f(y)(dy).
Now remember the SDE which is associated with the innitesimal generator A:
dX
t
=b(X
t
)dt +(X
t
)dB
t
,
X
0
=x
0
Often in what follows we shall want to consider the conditional expectation and probability
measures, conditional on x
0
= x, these will be denoted E
x
and P
x
respectively.
Theorem (Dirichlet Solution).
Dene a stopping time via
T := inft 0 : X
t
/ .
Then u(x) given by
u(x) := E
x
_
_
T
0
(X
s
)ds +f(X
T
)
_
,
solves the Dirichlet problem for f.
Proof
Dene
M
t
:= u(X
Tt
) +
_
tT
0
(X
s
)ds.
We shall now show that this M
t
is a martingale. For t T, it is clear that dM
t
= 0. For
t < T by Itos formula
dM
t
= du(X
t
) +(X
t
)dt.
Also, by Itos formula,
du(X
t
) =
d
j=1
u
x
i
(X
t
)dX
i
t
+
1
2
d
i=1
d
j=1
2
u
x
i
x
j
(X
t
)dX
i
, X
j
)
t
=
d
j=1
u
x
j
(X
t
) [b(X
t
)dt +(X
t
)dB
t
] +
1
2
d
i=1
d
j=1
d
k=1
ik
kj
d
u
x
i
x
j
(X
t
)dt
=Au(X
t
)dt +
d
j=1
(X
t
)
u
x
j
(X
t
)dB
t
.
Putting these two applications of Itos formula together yields
dM
t
= (Au(X
t
) +(X
t
)) dt +
d
j=1
(X
t
)
u
x
j
(X
t
)dB
t
.
Relations to Second Order PDEs 64
but since u solves the Dirichlet problem, then
(Au +)(X
t
) = 0,
hence
dM
t
=(Au(X
t
) +(X
t
)) dt +
d
j=1
(X
t
)
u
x
j
(X
t
)dB
j
t
,
=
d
j=1
(X
t
)
u
x
j
(X
t
)dB
j
t
.
from which we conclude by the stability property of the stochastic integral that M
t
is a
local martingale. However M
t
is uniformly bounded on [0, t], and hence M
t
is a martingale.
In particular, let (x) 1, and f 0, by the optional stopping theorem, since T t
is a bounded stoppping time, this gives
u(x) = E
x
(M
0
) = E
x
(M
Tt
) = E
x
[u(X
Tt
) + (T t)].
Letting t , we have via monotone convergence that E
x
(T) < , since we know that
the solutions u is bounded from the PDE solution existance theorem; hence T < a.s..
We cannot simply apply the optional stopping theorem directly, since T is not necessarily
a bounded stopping time. But for arbitrary and f, we have that
[M
t
[ u
+T
= sup
x
[u(x)[ +T sup
x
[(x)[,
whence as E
x
(T) < , the martingale M is uniformly integrable, and by the martingale
convergence theorem has a limit M
= f(X
T
) +
_
T
0
(X
s
)ds.
Hence from the identity E
x
M
0
= E
x
M
we have that,
u(x) = E
x
(M
0
) = E
x
(M
) = E
x
_
f(X
T
) +
_
T
0
(X
s
)ds
_
.
Relations to Second Order PDEs 65
17.3. The Cauchy Problem
The Cauchy Problem for f, a C
2
b
function, is the solution of the system
u
t
= Au on
u(0, x) = f(x) on x
u(t, x) = f(x) t 0, on x
A typical problem of this sort is to solve the heat equation,
u
t
=
1
2
2
u,
where the function u represents the temperature in a region , and the boundary condition
is to specify the temperature eld over the region at time zero, i.e. a condition of the form
u(0, x) = f(x) for x ,
In addition the boundary has its temperature xed at zero,
u(0, x) = 0 for x .
If is just the real line, then the solution has the beautifully simple form
u(t, x) = E
x
(f(B
t
)) ,
where B
t
is a standard Brownian Motion.
Theorem (Cauchy Existence) 17.3.
For each f C
2
b
(R
d
) there exists a unique u in C
1,2
b
(RR
d
) such that u solves the Cauchy
Problem for f. Moreover there exists a continuous function (the heat kernel)
p : (0, ) R
d
R
d
(0, ),
such that for all f C
2
b
(R
d
), the solution to the Cauchy Problem for f is given by
u(t, x) =
_
R
d
p(t, x, y)f(y)dy.
Theorem 17.4.
Let u C
1,2
b
(R R
d
) be the solution of the Cauchy Problem for f. Then dene
T := inft 0 : X
t
/ ,
a stopping time. Then
u(t, x) = E
x
[f(X
Tt
)]
Relations to Second Order PDEs 66
Proof
Fix s (0, ) and consider the time reversed process
M
t
:= u((s t) T, X
tT
).
There are three cases now to consider; for 0 T t s, M
t
= u((s t) T, X
T
),
where X
T
, so from the boundary condition, M
t
= f(X
T
), and hence it is clear that
dM
t
= 0. For 0 s T t and for 0 t s T, the argument is similar; in the latter
case by Itos formula we obtain
dM
t
=
u
t
(s t, X
t
)dt +
d
j=1
u
x
j
(s t, X
t
)dX
j
t
+
1
2
d
i=1
d
j=1
2
u
x
i
x
j
(s t, X
t
)dX
i
, X
j
)
t
,
=
_
u
t
+Au
_
(s t, X
t
)dt +
d
j=1
u
x
j
(s t, X
t
)
d
k=1
jk
(X
t
)dB
k
t
.
We obtain a similar result in the 0 t T s, case but with s replaced by T. Thus for
u solving the Cauchy Problem for f, we have that
_
u
t
+Au
_
= 0,
we see that M
t
is a local martingale. Boundedness implies that M
t
is a martingale, and
hence by optional stopping
u(s, x) = E
x
(M
0
) = E
x
(M
s
) = E
x
(f(X
sT
)),
17.4. FeynmanKac Representation
Feynman observed the following representation for the representation of the solution of a
PDE via the expectation of a suitable function of a Brownian Motion intuitively and the
theory was later made rigorous by Kac.
In what context was Feynman interested in this problem? Consider the Schrodinger Equa
tion,
2
2m
2
(x, t) +V (x)(x, t) = i
t
(x, t),
which is a second order PDE. Feynman introduced the concept of a pathintegral to express
solutions to such an equation. In a manner which is analogous to the Hamiltonian principle
in classical mechanics, there is an action integral which is minimised over all permissible
paths that the system can take.
Relations to Second Order PDEs 67
We have already considered solving the Cauchy problem
u
t
= Au on
u(0, x) = f(x) on x
u(t, x) = f(x) t 0, on x
where A is the generator of an SDE and hence of the form
A =
d
j=1
b
j
(X
t
)
x
j
+
1
2
d
i=1
d
j=1
a
ij
(X
t
)
2
x
i
x
j
.
Now consider the more general form of the same Cauchy problem where we consider a
Cauchy Problem with generator L of the form:
L A+v =
d
j=1
b
j
(X
t
)
x
j
+
1
2
d
i=1
d
j=1
a
ij
(X
t
)
2
x
i
x
j
+v(X
t
).
For example let X
t
= B
t
corresponding to
A =
1
2
2
, L =
1
2
2
+v(X
t
),
so in this example we are solving the problem
u
t
=
1
2
2
u(t, X
t
) +v(X
t
)u(X
t
) on R
d
.
u(0, x) = f(x) on R
d
.
The FeynmanKac Representation Theorem expresses the solution of a general second
order PDE in terms of an expectation of a function of a Brownian Motion. To simplify
the statement of the result, we shall work on = R
d
, since this removes the problem of
considering the Brownian Motion hitting the boundary.
Theorem (FeynmanKac Representation).
Let u C
1,2
b
(R R
d
) be a solution of the Cauchy Problem with a generator of the above
form with initial condition f, and let X
t
be a solution of the SDE with innitessimal
generator A in R
d
starting at x. Then
u(t, x) = E
x
_
f(X
t
) exp
__
t
0
v(X
s
)ds
__
.
Proof
Note that the SDE for X
t
has the form
dX
t
= a(X
t
)dB
t
+b(X
t
)dt.
Relations to Second Order PDEs 68
Fix s (0, ) and apply Itos formula to
M
t
= u(s t, X
t
) exp
__
t
0
v(X
r
)dr
_
.
For notational convenience, let
E
t
= exp
__
t
0
v(X
r
)dr
_
,
whence by Itos formula dE
t
= E
t
v(X
t
)dt. For 0 t s, we have
dM
t
=
d
j=1
u
x
j
(s t, X
t
)E
t
dB
j
t
+
_
u
t
+Au +vu
_
(s t, X
t
)E
t
dt
=
d
j=1
u
x
j
(s t, X
t
)E
t
dB
j
t
.
Hence M
t
is a local martingale; since it it bounded, M
t
is a martingale and hence by
optional stopping
u(s, x) = E
x
(M
0
) = E
x
(M
s
) = E
x
(f(X
s
)E
s
) .
18. Stochastic Filtering
Engineers have for many years studied system problems arising in the design of control
systems to be placed between user inputs to control system parameters in order to achieve
a desired response. For example the connection between the ight controls on a modern
aircraft such as the Harrier Jump Jet and the aircraft control surfaces go via a complex
control system rather than simple wire cables as found in early aircraft.
Even a basic control system such as a central heating thermostat is based upon an
element of feedback observing the state of the system (room temperature) and adjusting
the system inputs (turning the boiler o or on) to keep the system state close to the desired
value (the temperature set on the dial).
In the example of a high performance aircraft, sudden movement of an aircraft control
surface communicated directly to the control surfaces might cause uncontrolled growing
oscillations which was not the response required and the control system would have to
modify the control surfaces to avoid this happening.
In the real world the behaviour of these systems is aected by extra random unknown
factors which can not be known; for example wind eects on the airframe, engine vibration.
Also all measurements have inherent random errors which cannot be eliminated completely.
Attempts to solve control problems in this context require that we have some idea how to
use the observations made to give a best possible estimate of the system state at a given
time.
In this basic introduction to stochastic nonlinear ltering we shall consider a system
described by a stochastic process called the signal process whose behaviour is governed
by a dynamical system containing a Brownian motion noise input. This is observed, to
gether with unavoidable observation noise. The goal of the stochastic ltering problem is
to determine the law of the signal given the observations. This must be formalised in the
language of stochastic calculus. The notation   will be used in this section to denote the
Euclidean norm in ddimensional space.
An important modern example of stochastic ltering is the use of GPS measurements
to correct an inertial navigation system (INS). An INS system has transducers which mea
sure acceleration. Starting from a known point in space the acceleration can be integrated
twice to compute a position. But since there are error in acceleration measurements this
position will gradually diverge from the true position. This is a phenomenon familiar to
anyone who has navigated with a map by deadreckoning (From the station I walk NE for
a mile, E for 2 miles where am I?).
The simple navigator checks his position as frequently as possible by other means (e.g.
observing the bearing to several landmarks, taking a starsight or by looking at a handheld
GPS receiver).
In a similar fashion we might try and correct our INS position estimates by using
periodic GPS observations; these GPS position measurements can not be made continu
ously and are subject to random errors; additionally the observer continues to move while
taking the readings. What is the optimal way to use these GPS measurements to correct
our INS estimate? The answer lies in the solution of a stochastic ltering problem. The
signal process is the true position and the observation process is the GPS position process.
[69]
Stochastic Filtering 70
18.1. Signal Process
Let X
t
, T
t
, t 0 be the signal process. It is dened to be the solution of the stochastic
dierential equation
dX
t
= f (t, X
t
)dt +(t, X
t
)dV
t
, (4)
where V is a d dimensional standard Brownian motion. The coecients satisfy the condi
tions
f (t, x) f (t, y) +(t, x) (t, y) kx y
f (t, x)
2
+(t, x)
2
k
2
(1 +x
2
),
which ensure that the equation for X has a unique solution.
18.2. Observation Process
The observation process satises the stochastic dierential equation
dY
t
= h(t, X
t
)dt + dW
t
, Y
0
= 0, (5)
where W
t
is an m dimensional standard Brownian motion independent of V
t
. The function
h taking values in R
m
satises a linear growth condition
h(t, x)
2
k(1 +x
2
).
A consequence of this condition is that for any T > 0,
E
_
X
t

2
_
C
_
1 +E
_
X
0

2
_
e
ct
, (6)
for t [0, T] with suitable constants C and c which may be functions of T. As a result of
this bound
E
_
_
T
0
h(s, X
s
)
2
ds
_
< .
Given this process a sequence of observation algebras may be dened
o
t
:= (Y
s
: 0 s t) .
These must be augmented in the usual fashion to give
t
=
o
t
^,
where ^ is the set of P null subsets of , and
:=
t0
t
.
Stochastic Filtering 71
18.3. The Filtering Problem
The above has set the scene, we have a real physical system whose state at time t is
represented by the vector X
t
. The system state is governed by an SDE with a noise term
representing random perturbations to the system. This is observed to give the observation
process Y
t
which includes new independent noise (represented by W
t
).
The ltering problem is to nd the conditional law of the signal process given the
observations to date, i.e. to nd
t
() := E((X
t
)[
t
) t [0, T]. (7)
The following proofs are in many cases made complex by the fact that we do not
assume that the functions f, h and are bounded.
18.4. Change of Measure
To solve the ltering problem Girsanovs theorem will be used to make a change of measure
to a new measure under which the observation process Y is a Brownian motion. Let
Z
t
:= exp
_
_
t
0
h
T
(s, X
s
)dW
s
1
2
_
t
0
h(s, X
s
)
2
ds
_
, (8)
be a real valued process.
Proposition 18.1.
The process Z
t
is a martingale with respect to the ltration T
t
, t 0.
Proof
The process Z
t
satises
Z
t
= 1
_
t
0
Z
s
h
T
(s, X
s
)dW
s
,
that is Z
t
solves the SDE
dZ
t
= Z
t
h
T
(t, X
t
)dW
t
,
Z
0
= 1.
Given this expression, Z
t
is a positive continuous local martingale and hence is a super
martingale. To prove that it is a true martingale we must prove in addition that it has
constant mean.
Let T
n
be a reducing sequence for the local martingale Z
t
, i.e. an increasing sequence
of stopping times tending to innity as n such that for each n, Z
T
n
is a genuine
martingale. By Fatous lemma, and the local martingale property
EZ
t
= E lim
n
Z
T
n
t
liminf EZ
T
n
t
= EZ
T
n
0
= 1,
so
EZ
t
1 t. (9)
This will be used as an upper bound in an application of the dominated convergence
theorem. By application of Itos formula to the function f dened by
f(x) =
x
1 +x
,
Stochastic Filtering 72
we obtain that
f(Z
t
) = f(Z
0
) +
_
t
0
f
(Z
s
)dZ
s
+
1
2
_
t
0
f
(Z
s
)dZ
s
, Z
s
),
Hence
Z
t
1 +Z
t
=
1
1 +
_
t
0
Z
s
h
T
(s, X
s
)
(1 +Z
s
)
2
dW
s
_
t
0
Z
2
s
h(s, X
s
)
2
(1 +Z
s
)
3
ds. (10)
Consider the term
_
t
0
Z
s
h
T
(s, X
s
)
(1 +Z
s
)
2
dW
s
,
clearly this a local martingale, since it is a stochastic integral. The next step in the proof
is explained in detail as it is one which occurs frequently. We wish to show that the above
stochastic integral is in fact a genuine martingale. From the earlier theory (for L
2
integrable
martingales) it suces to show that integrand is in L
2
(W). We therefore compute
_
_
_
_
Z
s
h
T
(s, X
s
)
(1 +Z
s
)
2
_
_
_
_
W
= E
_
_
t
0
_
_
_
_
Z
s
h
T
(s, X
s
)
(1 +Z
s
)
2
_
_
_
_
2
ds
_
,
with a view to showing that it is nite, whence the integrand must be in L
2
(W) and hence
the integral a genuine martingale. We note that since Z
s
0, that
1
(1 +Z
s
)
1, and
Z
2
s
(1 +Z
s
)
2
1
2
,
so,
_
_
_
_
Z
s
h
T
(s, X
s
)
(1 +Z
s
)
2
_
_
_
_
2
=
Z
2
s
(1 +Z
s
)
4
h(s, X
s
)
2
=
_
Z
2
s
(1 +Z
s
)
2
_
_
1
(1 +Z
s
)
2
_
h(s, X
s
)
2
2
h(s, X
s
)
2
.
Using this inequality we obtain
E
_
_
t
0
_
_
_
_
Z
s
h
T
(s, X
s
)
(1 +Z
s
)
2
_
_
_
_
2
ds
_
2
E
_
_
T
0
h(s, X
s
)
2
ds
_
,
however as a consequence of our linear growth condition on h, the last expectation is nite.
Therefore
_
t
0
Z
s
h
T
(s, X
s
)
(1 +Z
s
)
2
dW
s
is a genuine martingale, and hence taking the expectation of (10) we obtain
E
_
Z
t
1 +Z
t
_
=
1
1 +
E
__
t
0
Z
2
s
h(s, X
s
)
2
(1 +Z
s
)
3
ds
_
.
Stochastic Filtering 73
Consider the integrand on the right hand side; clearly it tends to zero as 0. But also
Z
2
s
h(s, X
s
)
2
(1 +Z
s
)
3
=
Z
s
(1 +Z
s
)
3
Z
s
h(s, X
s
)
2
1 +Z
s
(1 +Z
s
)
3
Z
s
h(s, X
s
)
2
kZ
s
_
1 +X
s

2
_
,
where we have used the fact that h(s, X
s
)
2
k(1 + X
s

2
). So from the fact that
EZ
s
1 and using the next lemma we shall see that E
_
Z
s
X
s

2
_
C. Hence we conclude
that kZ
s
(1 + X
s

2
) is an integrable dominating function for the integrand on interest.
Hence by the Dominated Convergence Theorem, as 0,
E
__
t
0
Z
2
s
h(s, X
s
)
2
(1 +Z
s
)
3
ds
_
0.
In addition, since E(Z
s
) 1, the bounded convergence theorem yields
E
_
Z
s
1 +Z
s
_
E(Z
s
), as 0.
Hence we conclude that
E(Z
t
) = 1 t,
and so Z
t
is a genuine martingale.
Lemma 18.2.
E
__
t
0
Z
s
X
s

2
ds
_
< C(T) t [0, T].
Proof
To establish this result Itos formula is used to derive two important results
d(X
t

2
) =2X
T
t
(f dt +dW
s
) + tr(
T
)dt
d(Z
t
X
t

2
) =Z
t
X
t

2
h
T
dV
s
+Z
t
_
2X
T
t
(f dt +dW
s
) + tr(
T
)
_
dt
By Itos formula,
d
_
Z
t
X
t

2
1 +Z
t
X
t

2
_
=
1
(1 +Z
t
X
t

2
)
2
d
_
Z
t
X
t

2
_
+
(1 +Z
t
X
t

2
)
3
dZ
t
X
t

2
).
Substituting for d(Z
t
X
t

2
) into this expression yields
d
_
Z
t
X
t

2
1 +Z
t
X
t

2
_
=
1
(1 +Z
t
X
t

2
)
2
_
Z
t
X
t

2
h
T
dW
t
+Z
t
2X
T
t
dV
s
+
_
Z
t
_
2X
T
t
f + tr(
T
)
_
(1 +Z
t
X
t

2
)
2
_
Z
2
t
X
t

4
h
T
h + 4Z
2
t
2X
T
t
T
X
t
_
(1 +Z
t
X
t

2
)
3
_
dt.
Stochastic Filtering 74
After integrating this expression from 0 to t, the terms which are stochastic integrals
are local martingales. In fact they can be shown to be genuine martingales. For example
consider the term
_
t
0
Z
s
2X
T
s
(1 +Z
s
X
s

2
)
2
dV
s
,
to show that this is a martingale we must therefore establish that
E
_
_
_
t
0
_
Z
s
2X
T
s
(1 +Z
s
X
s

2
)
2
_
2
ds
_
_
= 4E
_
_
t
0
Z
2
s
X
T
s
T
X
s
(1 +Z
s
X
s

2
)
4
ds
_
< .
In order to establish this inequality consider the term X
T
t
T
X
t
which is a sum over terms
of the form
[X
i
t
ij
(t, X
t
)
kj
(t, X
t
)X
k
t
[ X
t

2

2
,
of which there are d
3
terms (in R
d
). But by the linear increase condition on , we have for
some constant

2
(1 +X
2
),
and hence
[X
i
t
ij
(t, X
t
)
kj
(t, X
t
)X
k
t
[ X
t

2
_
1 +X
t

2
_
,
so the integral may be bounded by
_
t
0
Z
2
s
X
T
s
T
X
s
(1 +Z
s
X
s

2
)
4
ds d
3
_
t
0
Z
2
s
X
s

2
_
1 +X
s

2
_
(1 +Z
s
X
s

2
)
4
ds
= d
3
_
_
t
0
Z
2
s
X
s

2
(1 +Z
s
X
s

2
)
4
ds +
_
t
0
Z
2
s
X
s

4
(1 +Z
s
X
s

2
)
4
_
ds
Considering each term separately, the rst satises
_
t
0
Z
2
s
X
s

2
(1 +Z
s
X
s

2
)
4
ds
_
t
0
Z
s
Z
s
X
s

2
(1 +Z
s
X
t

2
)
1
(1 +Z
s
X
s

2
)
3
ds
_
t
0
Z
s
ds
1
_
t
0
Z
s
.
The last integral has a bounded expectation since E(Z
s
) 1. Similarly for the second
term,
_
t
0
Z
2
s
X
s

4
(1 +Z
s
X
s

2
)
4
ds
_
t
0
Z
2
s
X
s

4
(1 +Z
s
X
s

2
)
2
1
(1 +Z
s
X
s

2
)
2
ds
1
2
t < .
A similar argument holds for the other stochastic integrals, so they must both be genuine
martingales, and hence if we take the expectations of the integrals they are zero. Integrating
Stochastic Filtering 75
the whole expression from 0 to t, then taking the expectation and nally dierentiation
with respect to t yields
d
dt
E
_
Z
t
X
t

2
1 +Z
t
X
t

2
_
E
_
Z
t
(2X
T
t
f + tr(
T
)
(1 +Z
t
X
t

2
)
2
_
k
_
1 +E
_
Z
t
X
t

2
1 +Z
t
X
t

2
__
.
An application of Gronwalls inequality (see the next section for a proof of this useful
result) establishes,
E
_
Z
t
X
t

2
1 +Z
t
X
t

2
_
C(T); t [0, T].
Applying Fatous lemma the desired result is obtained.
Given that Z
t
has been shown to be a martingale a new probability measure
P may
be dened by the RadonNikodym derivative:
d
P
dP
F
o
T
= Z
T
.
This change of measure allows us to establish one of the most important results in stochastic
ltering theory.
Proposition 18.3.
Under the probability measure
P, the observations process Y
t
is a Brownian motion, which
is independent of X
t
.
Proof
Dene a process M
t
taking values in R by
M
t
=
_
t
0
h
T
(s, X
s
) dW
s
, (11)
is a martingale. Since W
t
is a P Brownian Motion, which is a martingale under P, it follows
that M
t
is a P local martingale. The previous lemma then establishes that M
t
is in fact a
P martingale. Thus we may apply Girsanovs theorem, to construct a new measure where
Y
t
=
d
P
dP
t
= c(M
t
) = exp
_
_
t
0
h
T
(s, X
s
) dW
s
1
2
_
t
0
h(s, X
s
)
2
ds
_
= Z
t
.
As a corollary of Girsanovs theorem W
t
W, M)
t
is a continuous
P local martingale.
Levys characterisation theorem then implies that under
P the process
W
t
W, M)
t
= W
t
+
_
t
0
h(s, X
s
) ds,
Stochastic Filtering 76
is an m dimensional Brownian Motion. But this process is just the observation process Y
t
by denition. Thus we have proven that under
P, the observation process Y
t
is a Brownian
Motion. To establish the second part of the proposition we must prove the independence
of Y and X. The law of the pair (X, Y) on [0, T] is absolutely continuous with respect to
that of the pair (X, V) on [0, T], since the latter is equal to the former plus a drift term.
Their RadonNikodym derivative (which exists since they are absolutely continuous
with respect to each other) is
T
, which means that for f any bounded measurable function
E
P
[f(X
t
, Y
t
)
] = E
P
[E
P
[f(X
t
, Y
t
)
t
[T
t
]]
= E
P
[f(X
t
, Y
t
)E
P
[
[T
t
]]
= E
P
[f(X
t
, Y
t
)
t
] = E
P
[f(X
t
, V
t
)] .
Hence in terms of the probability measure
P
E
P
[f(X
t
, Y
t
)] = E
P
[f(X
t
, Y
t
)Z
] = E
P
[f(X
t
, V
t
)] ,
and hence under
P, the variables X and Y are independent.
Proposition 18.4.
Let U be an integrable T
t
measurable random variable. Then
E
P
[U[
t
] = E
P
[U[] .
Proof
Dene
t
:= (Y
t+u
Y
t
: u 0) ,
then =
t
t
. Under the measure
P, the algebra
t
is independent of T
t
, since Y is
an T
t
adapted Brownian Motion which implies that it satises the independent increment
property. Hence
E
P
[U[
t
] = E
P
[U[
t
t
] = E
P
[U[] .
18.5. The Unnormalised Conditional Distribution
To simplify the notation in what follows dene
Z
t
= 1/Z
t
. Then under the measure
P we
have
d
Z
t
=
Z
t
h
T
(t, X
t
)dY
t
, (12)
so
Z
t
is a
P local martingale. Hence
Z
t
= exp
__
t
0
h
T
(s, X
s
)dY
t
1
2
_
t
0
h(s, X
s
)
2
ds
_
.
Stochastic Filtering 77
Also for any stopping time T, E
P
_
Z
T
_
= E
P
_
Z
T
Z
T
_
= 1, the constant mean condition
implies that
Z
T
is a UI
P martingale and thus by the martingale convergence theorem
Z
t
Z
in L
1
. So we may write
dP
d
=
Z
,
whence
dP
d
F
t
=
Z
t
, for all t 0.
Now dene, for every bounded measurable function , the unnormalised conditional
distribution of X via
t
() :=E
P
_
(X
t
)
Z
t
[
t
_
=E
P
_
(X
t
)
Z
t
[
_
, (13)
using the result of proposition 18.4.
The following result within the context of stochastic ltering is sometimes called the
KallianpurStriebel formula.
Proposition 18.5.
For every bounded measurable function , we have
t
() := E
P
[(X
t
)[
t
] =
t
()
t
(1)
, P and
P a.s..
Proof
Let b be a bounded
t
measurable function. From the denition of
t
viz
t
() := E
P
[(X
t
)[
t
] ,
we deduce from the denition of conditional expectation that
E
P
(
t
()b) = E
P
((X
t
)b) ,
hence by the tower property of conditional expectation
E
P
_
Z
t
t
()b
_
= E
P
_
Z
t
(X
t
)b
_
.
Whence
E
P
_
t
()E
P
_
Z
t
[
t
_
b
_
= E
P
_
E
P
_
(X
t
)
Z
t
[
t
_
b
_
.
which is equivalent to
E
P
(
t
()
t
(1)b) = E
P
(
t
()b) .
This implies that
E
P
(
t
()
t
(1)[
t
) = E
P
(
t
()[
t
)
As the random variables
t
(),
t
() and
t
(1) from their denitions are all
t
measurable
this implies that
t
()
t
(1) =
t
()
Since
Z
t
> 0 a.s. which implies that
t
(1) > 0 a.s. this is sucient to prove the required
result.
Stochastic Filtering 78
Remark
The above proof is simply a specic proof of the abstract form of Bayes theorem, the
general form of which is a standard result, see e.g. A.0.3 in [Musiela and Rutkowski, 2005].
18.6. The Zakai Equation
We wish to derive a stochastic dierential equation which is satised by the unnormalised
density
t
.
Denition 18.6.
A set S is said to be total in L
1
(, T
t
, P), if for every a L
1
(, T
t
, P)
E(a
t
) = 0 S a = 0.
Lemma 18.7.
Let
S
t
:=
_
t
= exp
_
i
_
t
0
r
T
s
dY
s
+
1
2
_
t
0
r
s

2
ds
_
: r
s
L
([0, t], R
m
)
_
,
then the set S
t
is a total set in L
1
(,
t
,
P) and
E(a
t
) = 0 for all
t
S
t
that a = 0. Take
t
1
, t
2
, . . . , t
p
(0, t) such that t
1
< t
2
< < t
p
, then given k
1
, . . . , k
p
p
j=1
k
T
j
Y
t
j
=
p
j=1
T
i
(Y
t
j
Y
t
j1
) =
_
t
0
b
T
s
dY
s
where
p
= k
p
,
p1
= k
p
+k
p1
, . . . ,
1
= k
p
+ k
1
.
and
b
s
=
_
j
for t (t
j1
, t
j
)
0 for t (t
p
, T)
Hence
E
_
a exp
_
i
_
t
0
b
s
dY
s
__
=
E
_
_
a exp
_
_
i
p
j=1
k
T
j
Y
t
j
_
_
_
_
= 0
and the same holds for linear combinations by the linearity of
E viz
E
_
_
a
K
k=1
c
k
exp
_
_
i
p
j=1
k
T
j,k
Y
t
j
_
_
_
_
= 0
Stochastic Filtering 79
where c
1
, . . . , c
K
are complex coecients. If F(x
1
, . . . , x
p
) is a continuous bounded complex
valued function dened on (R
m
)
p
then since the set
_
_
_
f(x
1
, . . . , x
p
) =
K
k=1
c
k
exp
_
_
i
p
j=1
k
T
j,k
x
j
_
_
: c
k
C k, k
j,k
C j, k
_
_
_
is closed under complex conjugation, and separates points in (R
m
)
p
, the StoneWeierstrass
approximation theorem for complex valued functions implies that there exists a uniformly
bounded sequence of functions of the form
P
(n)
(x
1
, . . . , x
p
) =
K
(n)
k=1
c
(n)
k
exp
_
_
i
p
j=1
_
k
(n)
j,k
_
T
x
j
_
_
such that
lim
n
P
(n)
(x
1
, . . . , x
p
) = F(x
1
, . . . , x
p
)
hence
E
_
aF(Y
t
1
, . . . , Y
t
p
)
_
= 0
for every continuous bounded F and by the usual approximation argument this extends
to every F bounded Borel measurable with respect to (Y
t
1
, . . . , Y
t
p
). As 0 < t
1
< . . . <
t
p
< t were arbitrary it follows that
E(af) = 0 for every f bounded
t
measurable; taking
f(x) = a(x) n for arbitrary n then implies
E(a
2
n) = 0, whence a = 0
P a.s..
Remark
As a result if L
1
(,
t
,
t0
be a continuous T
t
measurable process such that
E
_
_
T
0
U
2
t
dt
_
< , T 0,
then for any t 0, and for j = 1, . . . , m the following holds
E
__
t
0
U
s
dY
j
s
t
_
=
_
t
0
E(U
s
[
s
) dY
j
s
.
Proof
As S
t
is a total set in L
1
(,
t
,
E
_
t
_
t
0
U
s
dY
j
s
_
=
E
_
t
_
t
0
E(U
s
[
s
) dY
j
s
_
.
Stochastic Filtering 80
To this end note that for every
t
S
t
the following holds
t
= 1 +
_
t
0
i
s
r
T
s
dY
s
,
and hence
E
_
t
_
t
0
U
s
dY
j
s
_
=
E
__
1 +
_
t
0
i
s
r
T
s
dY
s
___
t
0
U
s
dY
j
s
__
=
E
__
t
0
U
s
dY
j
s
_
+
E
__
t
0
i
s
r
j
s
U
s
ds
_
.
Here the last term is computed by recalling the denition of the covariation process, in
conjunction with the Kunita Watanabe identity. The rst term on the right hand side above
is a martingale (from the boundedness condition on U
s
) and hence its expectation vanishes.
Thus using the fact that the lebesgue integral and the expectation
E([) commute
E
_
t
_
t
0
U
s
dY
j
s
_
=
E
__
t
0
i
s
r
j
s
U
s
ds
_
=
E
_
E
__
t
0
i
s
r
j
s
U
s
ds
__
=
E
__
t
0
i
E
_
s
r
j
s
U
s
_
ds
_
=
E
__
t
0
i
s
r
j
s
E(U
s
[) ds
_
=
E
__
t
0
i
s
r
T
s
dY
s
_
t
0
E(U
s
[) dY
j
s
_
=
E
_
t
_
t
0
E(U
s
[) dY
j
s
_
.
But from proposition 18.4, it follows that since U
s
is
s
measurable
E(U
s
[) =
E(U
s
[
s
)
whence
E
_
t
_
t
0
U
s
dY
j
s
_
=
E
_
t
_
t
0
E(U
s
[
s
) dY
j
s
_
.
As this holds for all
t
in S
t
, and the latter is a total set, by the earlier remarks, this
establishes the result.
Lemma 18.9.
Let U
t
t0
be a real valued T
t
adapted continuous process such that
E
_
_
T
0
U
2
s
ds
_
< , T 0, (14)
Stochastic Filtering 81
and let R
t
be another continuous T
t
adapted process such that R, Y )
t
= 0 for all t. Then
E
__
t
0
U
s
dR
s
[
t
_
= 0.
Proof
As before use the fact that S
t
is a total set, so it suces to show that for each
t
S
t
E
_
t
_
t
0
U
s
dR
s
_
= 0.
As in the previous proof we note that each
t
in S
t
satises
t
= 1 +
_
t
0
i
s
r
T
s
dY
s
Hence substituting this into the above expression yields
E
_
t
_
t
0
U
s
dR
s
_
=
E
__
t
0
U
s
dR
s
_
+
E
___
t
0
i
s
r
T
s
dY
s
___
t
0
U
s
dR
s
__
=
E
__
t
0
U
s
dR
s
_
+
E
__
t
0
i
s
U
s
r
T
s
dY, R)
s
_
=
E
__
t
0
U
s
dR
s
_
=0.
The last term vanishes, because by the condition 14, the stochastic integral is a genuine
martingale. The other term vanishes because from the hypotheses Y, R)
t
= 0.
Remark
In the context of stochastic ltering a natural application of this lemma will be made by
setting R
t
= V
t
, the stochastic noise process driving the signal process.
We are now in a position to state and prove the Zakai equation. The Zakai equa
tion is important because it is a parabolic stochastic partial dierential equation which is
satised by
t
, and indeed its solution provides a practical means of computing
t
() =
E
_
(X
t
)
Z
t
[
t
_
. Indeed the Zakai equation provides a method by which numerical solu
tion of the nonlinear ltering problem may be approached by using recursive algorithms
to solve the stochastic dierential equation.
Theorem 18.10.
Let A be the innitesimal generator of the signal process X
t
, let the domain of this in
nitesimal generator be denoted T(A). Then subject to the usual conditions on f, and
h, the unnormalised conditional distribution on X satises the Zakai equation which is
t
() =
0
() +
_
t
0
s
(A
s
)ds +
_
t
0
s
(h
T
)dY
s
, t 0, T(A). (15)
Stochastic Filtering 82
Proof
We approximate
Z
t
by
t
=
Z
t
1 +
Z
t
,
noting that since Z
t
0, this approximation is bounded by
t
4
27
2
.
Thus we have approximated the process Z
t
by a bounded process. Application of Itos
formula to f(x) = x/(1 +x), yields
d
Z
t
= df(
Z
t
) = (1 +
Z
t
)
2
Z
t
h
T
dY
t
(1 +
Z
t
)
3
Z
2
t
h(t, X
t
)
2
.
The integration by parts formula yields,
d
_
t
(X
t
)
_
= (X
t
)d
Z
t
+
Z
t
d ((X
t
)) +
_
t
, (X
t
)
_
.
However from the denition of an innitesimal generator, we know that
(X
t
) (X
0
) =
_
t
0
A
s
(X
s
)ds +M
t
,
where M
t
is a martingale given by
M
t
=
d
j=1
_
t
0
d
i=1
x
i
(X
t
)
ij
(t, X
t
) dW
j
t
.
Using the expression for dZ
t
found earlier we may write
d
_
t
(X
t
)
_
=
_
t
A(X
t
) (X
t
)(1 +
Z
t
)
3
Z
2
t
h(t, X
t
)
2
_
dt
+
Z
t
dM
t
+(X
t
)(1 +
Z
t
)
2
Z
t
h
T
(t, X
t
)dY
t
In terms of the functions already dened we can write the innitesimal generator for X
t
as
A
s
(x) =
d
i=1
f
i
(s, x)
x
i
(x) +
1
2
d
i=1
d
j=1
_
d
k=1
ik
(s, x)
kj
(s, x)
_
2
x
i
x
j
(x),
Now we wish to compute
t
() =
E
_
t
(X
t
)[
t
_
Stochastic Filtering 83
we shall do this using the previous expression which on integration gives us
t
(X
t
) =
Z
0
(X
0
) +
_
t
0
_
s
A(X
s
) (X
t
)(1 +
Z
s
)
3
Z
2
s
h(s, X
s
)
2
_
ds
+
_
t
0
s
dM
s
+
_
t
0
(X
s
)(1 +
Z
s
)
2
Z
s
h
T
(s, X
s
)dY
s
(16)
Note that
E
_
Z
0
(X
0
)
t
_
=
E
_
1
1 +
(X
0
)
0
_
=
0
()
1
1 +
.
Hence taking conditional expectations of both sides of (16)
E
_
Z
t
(X
t
)
t
_
=
0
()
1 +
+
E
__
t
0
(X
s
)(1 +
Z
s
)
2
Z
s
h
T
(s, X
s
) dY
s
t
_
+
E
__
t
0
_
s
A(X
s
) (X
s
)(1 +
Z
s
)
3
Z
2
s
h(s, X
s
)
2
_
ds
t
_
+
E
__
t
0
s
dM
t
_
Applying lemma 18.8 since
Z
t
is bounded allows us to interchange the (stochastic) integrals
and the conditional expectations to give
E
_
Z
t
(X
t
)
t
_
=
0
()
1 +
+
_
t
0
E
_
(X
s
)(1 +
Z
s
)
2
Z
s
h
T
(s, X
s
)
t
_
dY
s
+
_
t
0
E
_
Z
s
A(X
s
) (X
s
)(1 +
Z
s
)
3
Z
2
s
h(s, X
s
)
2
t
_
ds
+
_
t
0
E
_
Z
t
_
dM
s
Now we take the 0 limit. Clearly
Z
t
Z
t
pointwise, and thus by the monotone
convergence theorem since Z
t
Z
t
as 0,
E
_
t
(X
t
)[
t
_
E
_
Z
t
(X
t
)[
t
_
=
t
(),
and also by the monotone convergence theorem
E
_
s
A
s
(X
s
)[
s
_
s
(A
s
).
We can bound this by
> 0,
E(Z
s
A
s
(X
s
)[
s
) A
s

E
_
Z
s
[
s
_
,
Stochastic Filtering 84
and the right hand side is an L
1
bounded quantity. Hence by the dominated convergence
theorem we have
_
t
0
E
_
s
A
s
(X
s
)[
s
_
ds
_
t
0
s
(A
s
) ds a.s.
Also note that
lim
0
(X
s
)
_
s
_
2
_
1 +
Z
s
_
1
h(s, X
s
)
2
= 0.
and since the function may be dominated by
(X
s
)
_
s
_
2
_
1 +
Z
s
_
1
h(s, X
s
)
2

2
Z
s
1 +
Z
s
Z
s
1 +
Z
s
1
1 +
Z
s
h(s, X
s
)
2
Z
s
h(s, X
s
)
2
Z
s
k(1 +X
s

2
)
But
E(
Z
s
k(1 +X
s

2
)) = kE
_
Z
s
Z
s
_
1 +X
s

2
_
_
= k
_
1 +E(X
s

2
)
_
k
_
1 +C
_
1 +E(X
0

2
)
_
e
ct
_
(17)
Where the nal inequality follows by (6). As a consequence of these two results, as the right
hand side of the above inequality is integrable, by the conditional form of the Dominated
Convergence theorem it follows that
E
_
(X
s
)
_
s
_
2
_
1 +
Z
s
_
1
h(s, X
s
)
2
_
0
as 0. By proposition 18.4 the sigma eld may be replaced by
t
which yields that
as 0
E
_
(X
s
)
_
s
_
2
_
1 +
Z
s
_
1
h(s, X
s
)
2
s
_
0.
Now note again that the dominating function (17) is also Lebesgue integrable and bounded
thus a second application of the Dominated convergence theorem yields
_
t
0
E
_
(X
s
)
_
s
_
2
_
1 +
Z
s
_
1
h(s, X
s
)
2
s
_
ds 0, as 0.
It now remains to check that the stochastic integral
_
t
0
E
_
(X
s
)
Z
s
_
1 +
Z
s
_
1
h
T
(s, X
s
)
s
_
dY
s
_
t
0
s
(h
T
(X
s
)) dY
s
.
Stochastic Filtering 85
To this end dene
I
(t) :=
_
t
0
E
_
(X
s
)
Z
s
_
1 +
Z
s
_
1
h
T
(s, X
s
)
s
_
dY
s
,
and the desired limit
I(t) :=
_
t
0
s
(h
T
(X
s
))dY
s
=
_
t
0
E
_
Z
s
h
T
(s, X
s
)(X
s
)
t
_
dY
s
.
Both I(t) and I
(t) I(t) =
_
t
0
E
_
_
_
_
_
Z
s
_
1 +
Z
s
_
2
Z
s
_
_
_(X
s
)h
T
(s, X
s
)
_dY
s
=
_
t
0
E
_
Z
2
s
_
2 +
Z
s
_
_
1 +
Z
s
_
2
(X
s
)h
T
(s, X
s
)
_dY
s
,
Consider the L
2
norm of this dierence and use the fact that Y is a
P Brownian motion
E(I
(t) I(t))
2
=
E
_
_
_
_
t
0
E
_
Z
2
s
_
2 +
Z
s
_
_
1 +
Z
s
_
2
(X
s
)h
T
(s, X
s
)
_dY
s
_
_
_
2
=
E
_
_
_
_
t
0
_
_
_
_
_
_
_
E
_
Z
2
s
_
2 +
Z
s
_
_
1 +
Z
s
_
2
(X
s
)h
T
(s, X
s
)
_
_
_
_
_
_
_
_
2
ds
_
_
_
By the conditional form of Jensens inequality 
E(X[)
2
E
_
X
2
_
. Thus
_
_
_
_
_
_
_
E
_
Z
2
s
_
2 +
Z
s
_
_
1 +
Z
s
_
2
(X
s
)h
T
(s, X
s
)
_
_
_
_
_
_
_
_
2
E
_
_
_
_
_
_
_
_
_
_
Z
2
s
_
2 +
Z
s
_
_
1 +
Z
s
_
2
(X
s
)h
T
(s, X
s
)
_
_
_
_
_
_
_
2
_
_
_
A dominating function can can readily be found for the quantity inside the conditional
expectation viz
_
_
_
_
_
_
_
Z
2
s
_
2 +
Z
s
_
_
1 +
Z
s
_
2
(X
s
)h
T
(s, X
s
)
_
_
_
_
_
_
_
2

2
Z
s
1 +
Z
s
_
1 +
1
1 +
Z
s
_
Z
s
2
h(s, X
s
)
2
4
2
Z
2
s
h(s, X
s
)
2
4k
2
Z
2
s
_
1 +X
s

2
_
Stochastic Filtering 86
Since
E
_
Z
s
_
Z
s
+
Z
s
X
s

2
__
= E
_
Z
s
_
+E
_
Z
s
X
s

2
_
Using the dominated convergence theorem as 0
_
t
0
_
_
_
_
_
_
_
E
_
Z
2
s
_
2 +
Z
s
_
_
1 +
Z
s
_
2
(X
s
)h
T
(s, X
s
)
t
_
_
_
_
_
_
_
_
_
2
ds 0
Now we use the dominated convergence theorem to show that for a suitable subse
quence that since h(s, X
s
)
2
K(1 +X
s

2
) we can write
E
_
_
_
t
0
_
_
_
E
_
_(X
s
)
Z
2
s
_
2 +
Z
s
_
_
1 +
Z
s
_
2
h
T
(s, X
s
)[
s
_
_
_
_
_
2
ds
_
_ 0. a.s.
Thus we have shown that
E
_
(I
(t) I(t))
2
_
0.
By a standard theorem of convergence, this means that there is a suitable subsequence
n
such that
I
n
(t) I(t) a.s.
which is sucient to establish the claim.
18.7. KushnerStratonowich Equation
The Zakai equation which has been derived in the previous section is an SDE which is
satised by the unnormalised conditional distribution of X
t
, i.e.
t
(). It seems natural
therefore to derive a similar equation satised by the normalised conditional distribution,
t
(). Recall that the unnormalised conditional distribution is given by
t
() =
E
_
(X
t
)
Z
t
[
t
_
,
and the normalised conditional law is given by
t
() := E[(X
t
)[
t
] .
As a consequence of the proposition 17.5
t
() =
t
()
t
(1)
,
Now using this result together with the Zakai equation, we can derive the Kushner
Stratonovitch equation.
Theorem (KushnerStratonovich) 18.11.
The normalised conditional law of the process X
t
subject to the usual conditions on the
processes X
t
and Y
t
satises the KushnerStratonovich equation
t
() =
0
() +
_
t
0
s
(A
s
)ds +
_
t
0
_
s
(h
T
)
s
(h
T
)
s
()
_
(dY
s
s
(h)ds) . (18)
The process Y
t
_
t
0
s
(h(s, X
s
)) ds is called the innovations process.
19. Gronwalls Inequality
An important and frequently used result in the theory of stochastic dierential equations
is Gronwalls lemma. There are various versions of Gronwalls theorem (see [Ethier and
Kurtz, 1986], [Karatzas and Shreve, 1987]) but the following general form and proof follows
that given in [Mandelbaum et al., 1998].
Theorem (Gronwall).
Let x, y and z be measurable nonnegative functions on the real numbers. If y is bounded
and z is integrable on [0, T], and for all 0 t T then
x(t) z(t) +
_
t
0
x(s)y(s)ds, (19)
then
x(t) z(t) +
_
t
0
z(s)y(s) exp
__
t
s
y(r)dr
_
ds.
Proof
Multiplying both sides of the inequality (19) by y(t) exp
_
_
t
0
y(s)ds
_
yields
x(t)y(t) exp
_
_
t
0
y(s)ds
_
__
t
0
x(s)y(s)ds
_
y(t) exp
_
_
t
0
y(s)ds
_
z(t)y(t) exp
_
_
t
0
y(s)ds
_
.
The left hand side can be written as the derivative of a product,
d
dt
___
t
0
x(s)y(s)ds
_
exp
_
_
t
0
y(s)ds
__
z(t)y(t) exp
_
_
t
0
y(s)ds
_
.
This can now be integrated to give
__
t
0
x(s)y(s)ds
_
exp
_
_
t
0
y(s)ds
_
_
t
0
z(s)y(s) exp
_
_
s
0
y(r)dr
_
ds,
or equivalently
_
t
0
x(s)y(s)ds
_
t
0
z(s)y(s) exp
__
t
s
y(r)dr
_
ds.
Combining this with the original equation (19) gives the desired result.
.
[87]
Gronwalls Inequality 88
Corollary 19.1.
If x, y, and z satisfy the conditions for Gronwalls theorem then
sup
0tT
x(t)
_
sup
0tT
z(t)
_
exp
_
_
T
0
y(s)ds
_
.
Proof
From the conclusion of Gronwalls theorem we see that for all 0 t T
x(t) sup
0tT
z(t) +
_
T
0
z(s)y(s) exp
_
_
T
s
y(r)dr
_
ds,
which yields
sup
0tT
x(t) sup
0tT
z(t) +
_
T
0
z(s)y(s) exp
_
_
T
s
y(r)dr
_
ds,
sup
0tT
z(t) +
_
sup
0tT
z(t)
__
T
0
y(s) exp
_
_
T
s
y(r)dr
_
ds
sup
0tT
z(t) +
_
sup
0tT
z(t)
_
_
exp
_
_
T
0
y(s)ds
_
1
_
_
sup
0tT
z(t)
_
exp
_
_
T
0
y(s)ds
_
.
20. Kalman Filter
It is clear that in the general case the SDEs given by the KushnerStratonivich equation
are innite dimensional. A special case in which they are of nite dimension is the most
famous example of stochastic ltering is undoubtedly the KalmanBucy lter. The relevant
equations can be derived without using the general machinery of stochastic ltering theory.
The rst practical use of the Kalman lter was in the ight guidance computer for the
Apollo missions, long before the general nonlinear stochastic ltering theory had been
developed. However by examining this special linear case, it is hoped that the operation
of the KushnerStratonowich equation may be made clearer.
We consider the system (X
t
, Y
t
) R
N
R
M
where X
t
is the signal process and Y
t
is the observation process, described by the following SDEs,
dX
t
= (B
t
X
t
+b
t
) dt +F
t
dV
t
dY
t
= H
t
X
t
dt + dW
t
where B
t
and F
t
are a N N matrix valued processes, V
t
is a N dimensional Brownian
motion, H
t
is an M N matrix valued process and W
t
is an M dimensional Brownian
Motion, which is independent of V
t
.
Thus in terms of the notation used earlier
f(t, X
t
) = b
t
+B
t
X
t
(t, X
t
) = F
t
h(t, X
t
) = H
t
X
t
The innitesimal generator of the signal process X
t
is given by
A
t
u =
N
i=1
b
(i)
t
(X
t
)
u
x
i
+
1
2
N
i=1
N
j=1
(FF
T
)
ij
2
u
x
i
x
j
which may readily be veried by applying Itos formula to f(X
t
).
Since the system of equations is linear it is simple to show that the conditional dis
tribution of X
t
conditional upon
t
is a multivariate normal distribution. Thus it may
be characterized by its rst and second moments alone. It is therefore sucient to con
sider (X) = x
i
and (X) = x
i
x
j
as test functions in the KushnerStratonovich equation
(18). The notation in what follows will become somewhat messy, we write X
(i)
t
for the ith
component of the signal process.
Since we shall only be concerned with these two moments, it will simplify the notation
to dene the conditional mean process
X
(i)
t
:=
t
(x
i
) = E
_
X
(i)
t
t
_
and the conditional covariance process
R
ij
t
= E
__
X
(i)
t
X
(i)
t
__
X
(j)
t
X
(j)
t
__
=
t
(x
i
x
j
)
t
(x
i
)
t
(x
j
)
Thus at time t our best estimate of the signal will be a normal distribution N(
X
t
, R
t
).
[89]
Kalman Filter 90
20.1. Conditional Mean
To derive an equation for the evolution of the conditional mean process, we apply the
KushnerStratonowich equation to (X) = X
(i)
, from theorem 18.11 this satises
X
(i)
t
=
X
(i)
0
+
_
t
0
E
_
(B
s
X
s
)
(i)
+b
(i)
t
t
_
dt
+
_
t
0
_
E
_
X
(i)
s
(H
s
X
s
)
T
[
s
_
X
(i)
s
(H
s
X
s
)
T
_
dN
s
where N
t
, the innovations process, is
N
t
= Y
t
_
t
0
H
s
X
s
ds
Whence writing the equation for the evolution of the conditional mean in vector form, we
obtain (since R
T
= R)
d
X
t
=
_
B
t
X
t
+b
s
_
dt +R
t
H
T
t
(dY
t
H
t
X
t
dt) (20)
20.2. Conditional Covariance
We now derive an equation for the evolution of the covariance matrix R, but rst we shall
need a result about the third moment of a multivariate normal distribution.
If X N(
X, R) then
E
_
X
(i)
t
X
(j)
t
X
(k)
t
_
= E
_
(
X
(i)
t
+Z
(i)
)(
X
(j)
t
+Z
(j)
)(
X
(k)
t
+Z
(k)
t
)
_
=
X
(i)
t
X
(j)
t
X
(k)
t
+
X
(i)
t
R
jk
t
+
X
j
t
R
ik
t
+
X
(k)
t
R
ij
t
Applying the KushnerStratonovich equation to (X) = x
i
x
j
yields
E
_
X
(i)
t
X
(j)
t
t
_
= E
_
X
(i)
0
X
(j)
0
0
_
+
_
t
0
E
_
(F
t
F
T
t
)
ij
+ (B
t
X
t
+b)
(i)
X
(j)
t
+ (B
t
X
t
+b)
(j)
X
(i)
t
t
_
dt
+
_
t
0
_
E(X
(i)
t
X
(j)
t
H
t
X
t
[
t
) E(X
(i)
t
X
(j)
t
[
t
)E(H
t
X
t
[
t
)
_
dN
s
(21)
It is clear that
dR
ij
= d
_
E(X
(i)
t
X
(j)
t
[
t
)
_
d
_
X
(i)
t
X
(j)
t
_
(22)
the rst term is given by (21), and using Itos form of the product rule to expand out the
second term
d
_
X
(i)
t
X
(j)
t
_
=
X
(i)
d
X
(j)
+
X
(j)
d
X
(i)
+ d[
X
(i)
,
X
(j)
]
t
Kalman Filter 91
Therefore using (20) we can evaluate this as
d
_
X
(i)
t
X
(j)
t
_
=
X
(i)
t
(B
t
X
t
+b
t
)
(j)
dt +
X
(i)
t
(R
t
H
T
t
dN
t
)
(j)
+
X
(j)
t
(B
t
X
t
+b
t
)
(i)
dt
+
X
(j)
t
(R
t
H
T
t
dN
t
)
(i)
+
_
(R
t
H
T
t
dN
t
)
(i)
, (R
t
H
T
t
dN
t
)
(j)
_
(23)
For evaluating the quadratic covariation term in this expression it is simplest to work
componentwise using the Einstein summation convention and the fact that the innovation
process N
t
is readily shown by Levys characterization to be a P Brownian motion
_
(R
t
H
T
t
dN
t
)
(i)
, (R
t
H
T
t
dN
t
)
(j)
_
=
_
R
il
t
H
kl
dN
k
, R
jm
H
nm
dN
n
= R
il
H
kl
R
jm
H
nm
kn
dt
= R
il
H
kl
H
km
R
jm
dt
= (RH
T
HR
T
)
ij
dt
= (RH
T
HR)
ij
dt (24)
where the last equality follows since R
T
= R. Substituting (24) into (23) and then using
this and (21) in (22) yields the following equation for the evolution of the ijth element of
the covariance matrix
dR
ij
t
= dt
_
1
2
(F
t
F
T
t
)
ij
+ (B
t
R
t
)
ij
+ (R
T
t
B
T
t
)
ij
(RH
T
HR)
ij
_
+ (
X
(i)
t
R
jm
t
+
X
(j)
t
R
jm
t
)H
lm
dN
l
(
X
i
t
R
jm
t
+
X
j
t
R
jm
t
)H
lm
dN
l
Thus we obtain the nal dierential equation for the evolution of the conditional covariance
matrix; notice that all of the stochastic terms have cancelled
dR
t
dt
= F
t
F
T
t
+B
t
R
t
+R
t
B
T
t
R
t
H
T
t
H
t
R
t
(25)
The two equations (20) and (25) therefore provide a complete description of the best
prediction of the system state X
t
, conditional on the observations up to time t. This is the
KalmanBucy lter.
It is most frequently encountered in a discrete time version, since in this form it is
practical for use in some of the problems described in the introduction to the stochastic
ltering section of these notes.
21. Discontinuous Stochastic Calculus
So far we have discussed integrals with repect to continuous semimartingales. It might
seem that this includes a great deal of the cases which are of real interest, but a very
important class of processes have been neglected.
The standard Poisson process on the line should be familiar. This has trajectories p(t)
which are T
t
adapted and piecewise continuous. These have the properties that for any
t 0, p(t) is a.s. nite and for any t, s 0, p(t + s) p(t) is independent of T
t
. The
process is also stationary the distribution of p(t +s) p(t) does not depend upon t. If the
probability of a jump in an innitessimal time interval is proportional to the length of the
time interval, the constant of proportionality is called the intensity and it can be shown
that the inter jump times have exponential distribution with parameter (i.e. mean 1/).
Such processes can also be described by their jump measures. The jump measure
N(, B) is the number of jumps within the time interval in the (Borel) set B. From the
above denition for any B B,
E(N([0, t +s], B) N([0, t], B) [T
t
) = (B),
where is a probability measure. More formally, we make the following denition
Denition 1. Poisson Process
A Poisson process with intensity measure h(dt, dy) = dtdy as a measurable map N from
(, T, P) into the space of positive integer valued counting measures on (R
+
, B(R
+
))
with the properties
(i) For every t 0 and every Borel subset A of [0, t] , N(A) is T
t
measurable.
(ii) For every t 0 and every collection of Borel subsets A
i
of [t, ) , N(A
i
) is
independent of T
t
.
(iii) E(N(A)) = h(A) for every Borel subset A of R
+
.
(iv) For each s > 0 and Borel B the distribution of N([t, t + s], B) does not depend
upon t.
This measure can be related back to the path p() by the following integral for any
t 0,
p(t) =
_
t
0
_
N(ds, d).
The above integral can be dened without diculty!
21.1. Compensators
Let p(t) be an T
t
adapted jump process with bounded jumps, such that p(0) = 0 and
E(p(t)) < for all t 0. Let (t) be a bounded nonnegative T
t
adapted process with
RCLL sample paths. If
p(t)
_
t
0
(s) ds
is an T
t
martingale, then () is the intensity of the jump process p() and the process
_
t
0
(s) ds is the compensator of the process p().
[92]
Discontinuous Stochastic Calculus 93
Why are compensators interesting? Recall from the theory of stochastic integration
with respect to continuous semimartingales that one of the most useful properties was that
a stochastic integral with respect to a continuous local is a local martingale. We would
like the same sort of property to hold when we integrate with respect to a suitable class
of discontinuous process.
Denition 21.2.
For A a nondecreasing locally integrable process, there is a nondecreasing previsible
process, called the compensator of A, denoted A
p
which is a.s. unique characterised by one
of the following
(i) AA
p
is a local martinagle.
(ii) E(A
p
T
) = E(A
T
) for all stopping times T.
(iii) E[(H A
p
)
] = E[(H A)
i=1
f
x
i
(X
s
) dX
s
+
1
2
n
i,j=1
_
t
0
2
f
x
i
x
j
(X
s
) d[(M
c
)
i
, (M
c
)
j
]
s
+
0<st
_
f(X
s
) f(X
s
)
n
i=1
f
x
i
(X
s
)X
i
s
_
.
The rather ugly notation (M
c
)
i
refers to the ith component of the continuous local mar
tingale M
c
and X
s
= X
s
X
s
.
A rough heuristic explanation for this generalized form of Ito and an aid to memory
of the formula is obtained by noting that the rst three terms on the right hand side are
identical to those in the multidimensional form of Ito for continuous processes.
Also terms 1,2 and 4 are exactly what you would have expected by writing down an
analogue of Ito for discontinuous processes of nite variation. In other words if the process
X has a jump at time s, the value of f(X) changes by f(X
s
) f(X
s
). The jump in X
s
will give rise to a discontinuity in
n
i=1
_
t
0
f
x
i
(X
s
) dX
s
with the jump having size
n
i=1
f
x
i
(X
s
)(X
s
X
s
)
To get the correct jump of f(X
s
) f(X
s
) then the corrective term is
0<st
_
f(X
s
) f(X
s
)
n
i=1
f
x
i
(X
s
)X
i
s
_
.
21.2. RCLL processes revisited
We shall be working extensively in the space of right continuous left limits processes (CAD
LAG) and we shall need a metric on the function space. First let us consider the space
D
T
of CADLAG functions from [0, T] to R. The usual approach of using the sup norm on
[0, T] isnt going to be very useful. Consider two functions dened on [0, 10] by
f(x) =
_
0 for x < 1,
1 for x 1,
and
g(x) =
_
0 for x < 1 +,
1 for x 1 +.
Discontinuous Stochastic Calculus 95
For arbitrary small these functions are always one unit apart in the sup metric. We
attempt to surmount this problem by considering the class
T
of increasing maps from
[0, T] to [0, T]. Now dene a metric
d
T
(f, g) := inf
_
: sup
0sT
[s (s)[ , sup
0sT
[f(s) g((s))[ for some
T
_
.
This metric would appear to nicely have solved the problem described above. Unfortunately
it is not complete! This is a major problem because Prohorovs theorem (among others)
requires a complete separable metric space! However the problem can be xed via a simple
trick; for a time transform
T
dene a norm
[[ := sup
0sT
log
_
(t) (s)
t s
_
.
Thus we replace the condition [(s) s[ by [[ and we obtain a new metric on D
T
which this time is complete and separable;
d
T
(f, g) := inf
_
: [[ , sup
0sT
[f(s) g((s))[ , for some
T
_
.
This is dened on the space of CADLAG functions on the compact time interval [0, T]
and can be exteneded via another standard trick to D
,
d
(f, g) :=
_
0
e
t
min [1, d
t
(f, g)] dt.
22. References
Bensoussan, A. (1982). Stochastic control of partially observable systems. CUP.
Bollobas, B. (1990). Linear Analysis. Cambridge.
Dellacherie, C. and Meyer, P.A. (1975). Probabilites et Potentiel A. Hermann, Paris.
Dellacherie, C. and Meyer, P.A. (1980). Probabilites et Potentiel B. Hermann, Paris.
Durrett, R. (1996). Probability: Theory and Examples. Duxbury.
Ethier, S. and Kurtz, T. (1986). Markov Processes Characterization and Convergence.
Wiley.
Karatzas, I. and Shreve, S. E. (1987). Brownian Motion and Stochastic Calculus. Springer.
Mandelbaum, A., Massey, W. A., and Reiman, M. I. (1998). Strong approximations for
markovian service networks. Queueing Systems Theory and Applications, 30:149
201.
Musiela, M. and Rutkowski, M. (2005). Martingale Methods in Financial Modelling.
Springer, 2nd edition.
Pardoux, E. (1989). Filtrage non lineaire et equations aux derivees partielles stochastiques
associees, pages 67163. Number 1464 in Lecture notes in Mathematics. Springer
Verlag.
Protter, P. (1990). Stochastic Integration and Dierential Equations. Springer.
Rogers, L. C. G. and Williams, D. (1994). Diusions, Markov Processes and Martingales:
Volume One: Foundations. Wiley.
Rogers, L. C. G. and Williams, D. (2000). Diusions, Markov Processes and Martingales:
Volume Two: Ito Calculus. Cambridge University Press, 2nd edition.
[96]