Stochastic Differential Equations

Do not worry about your problems with mathematics, I assure you mine are far
greater.

Albert Einstein.

Florian Herzog

2010
 Stochastic Differential Equations (SDE)

A ordinary differential equation (ODE)

dx(t)
= f (t, x) , dx(t) = f (t, x)dt , (1)
dt
with initial conditions x(0) = x0 can be written in integral form
Z t
x(t) = x0 + f (s, x(s))ds , (2)
0

where x(t) = x(t, x0, t0) is the solution with initial conditions x(t0) = x0. An
example is given as
dx(t)
= a(t)x(t) , x(0) = x0 . (3)
dt

Stochastic Systems, 2010 2

 Stochastic Differential Equations (SDE)

When we take the ODE (3) and assume that a(t) is not a deterministic parameter
but rather a stochastic parameter, we get a stochastic differential equation (SDE). The
stochastic parameter a(t) is given as

a(t) = f (t) + h(t)ξ(t) , (4)

where ξ(t) denotes a white noise process.

Thus, we obtain
dX(t)
= f (t)X(t) + h(t)X(t)ξ(t) . (5)
dt
When we write (5) in the differential form and use dW (t) = ξ(t)dt, where dW (t)
denotes differential form of the Brownian motion,we obtain:

dX(t) = f (t)X(t)dt + h(t)X(t)dW (t) . (6)

Stochastic Systems, 2010 3

 Stochastic Differential Equations (SDE)

In general an SDE is given as

dX(t, ω) = f (t, X(t, ω))dt + g(t, X(t, ω))dW (t, ω) , (7)

where ω denotes that X = X(t, ω) is a random variable and possesses the initial
condition X(0, ω) = X0 with probability one. As an example we have already
encountered
dY (t, ω) = µ(t)dt + σ(t)dW (t, ω) .
Furthermore, f (t, X(t, ω)) ∈ R, g(t, X(t, ω)) ∈ R, and W (t, ω) ∈ R. Similar as
in (2) we may write (7) as integral equation
Z t Z t
X(t, ω) = X0 + f (s, X(s, ω))ds + g(s, X(s, ω))dW (s, ω) . (8)
0 0

Stochastic Systems, 2010 4

 Stochastic Integrals

RT
For the calculation of the stochastic integral 0 g(t, ω)dW (t, ω), we assume that
g(t, ω) is only changed at discrete time points ti (i = 1, 2, 3, ..., N − 1), where
0 = t0 < t1 < t2 < . . . < tN −1 < tN < T . We define the integral
Z T
S= g(t, ω)dW (t, ω) , (9)
0

as the Riemannßum
N
X ³ ´
SN (ω) = g(ti−1, ω) W (ti, ω) − W (ti−1, ω) . (10)
i=1

with N → ∞ .

Stochastic Systems, 2010 5

 Stochastic Integrals

A random variable S is called the Itô integral of a stochastic process g(t, ω) with
respect to the Brownian motion W (t, ω) on the interval [0, T ] if

h³ N
X ³ ´i
lim E S − g(ti−1, ω) W (ti, ω) − (W (ti−1, ω) = 0, (11)
N →∞
i=1

for each sequence of partitions (t0, t1, . . . , tN ) of the interval [0, T ] such that
maxi(ti − ti−1) → 0. The limit in the above definition converges to the stochastic
integral in the mean-square sense. Thus, the stochastic integral is a random variable,
the samples of which depend on the individual realizations of the paths W (., ω).

Stochastic Systems, 2010 6

 Stochastic Integrals

The simplest possible example is g(t) = c for all t. This is still a stochastic process, but
a simple one. Taking the definition, we actually get
Z T N ³
X ´
c dW (t, ω) = c lim W (ti, ω) − W (ti−1, ω)
0 N →∞
i=1

= c lim [(W (t1, ω)−W (t0, ω)) + (W (t2, ω)−W (t1, ω)) + . . .
N →∞

+(W (tN , ω)−W (tN −1, ω))

= c (W (T, ω) − W (0, ω)) ,

where W (T, ω) and W (0, ω) are standard Gaussian random variables. With W (0, ω) = 0,
the last result becomes Z T
c dW (t, ω) = c W (T, ω) .
0

Stochastic Systems, 2010 7

 Stochastic Integrals

RT
Example: g(t, ω) = W (t, ω) 0
W (t, ω) dW (t, ω) =

N
X ³ ´
= lim W (ti−1, ω) W (ti, ω) − W (ti−1, ω)
N →∞
i=1

h1 XN
1X
N i
2 2 2
= lim (W (ti, ω) − W (ti−1, ω)) − (W (ti, ω) − W (ti−1, ω))
N →∞ 2 2 i=1
i=1

X N
1 2 1 2
= − lim (W (ti, ω) − W (ti−1, ω)) + W (T, ω) , (12)
2 N →∞ i=1 2

where we have used the following algebraic relationship y(x − y) = yx − y 2 + 12 x2 − 12 x2 =

1 2 1 2 1 2
2 x − 2 y − 2 (x − y) .

Stochastic Systems, 2010 8

 Stochastic Integrals

PN
We take now a detailed look at :limN →∞ i=1 (W (ti , ω) − W (ti−1, ω))2.
N
X N
X
2 2
E[ lim (W (ti, ω) − W (ti−1, ω)) ] = lim E[(W (ti, ω) − W (ti−1, ω)) ]
N →∞ N →∞
i=1 i=1

N
X
= lim (ti − ti−1)
N →∞
i=1

= T
N
X N
X
2 2
Var[ lim (W (ti, ω) − W (ti−1, ω)) ] = lim Var[(W (ti, ω) − W (ti−1, ω)) ]
N →∞ N →∞
i=1 i=1

N
X 2
= 2 lim (ti − ti−1) .
N →∞
i=1

Stochastic Systems, 2010 9

 Stochastic Integrals

By reducing the partition, the variance becomes zero,

N
X N
X
2
lim (ti − ti−1) ≤ max(ti − ti−1) lim (ti − ti−1)
N →∞ i N →∞
i=1 i=1

= max(ti − ti−1) T
i

= 0, (13)
PN 2
since ti−1 − ti → 0. Since the expected value of i=1 (ti − ti−1 ) is T and the
variance becomes zero, we get

N
X 2
(W (ti, ω) − W (ti−1, ω)) = T (14)
i=1

Stochastic Systems, 2010 10

 Stochastic Integrals

The stochastic integral has the solution

Z T
1 2 1
W (t, ω) dW (t, ω) = W (T, ω) − T (15)
0 2 2

This is inRcontrast to our intuition from standard calculus. In the case of a deterministic
T
integral 0 x(t)dx(t) = 12 x2(t), whereas the Itô integral differs by the term − 12 T .
— This example shows that the rules of differentiation (in particular the chain rule)
and integration need to be re-formulated in the stochastic calculus.

Stochastic Systems, 2010 11

 Stochastic Integrals

Properties of Itô Integrals.

Z T
E[ g(t, ω) dW (t, ω)] = 0 .
0

Proof:
Z T N
X ³ ´
E[ g(t, ω)dW (t, ω)] = E[ lim g(ti−1, ω) W (ti, ω) − W (ti−1, ω) ]
0 N →∞
i=1

N
X ³ ´
= lim E[g(ti−1, ω)] E[ W (ti, ω) − W (ti−1, ω) ]
N →∞
i=1

= 0.

The expectation of stochastic integrals is zero. This is what we would expect anyway.

Stochastic Systems, 2010 12

 Stochastic Integrals

Properties of Itô Integrals.

hZ T i Z T
2
Var g(t, ω)dW (t, ω) = E[g (t, ω)]dt . (16)
0 0

Proof:
hZ T i h Z T i
2
Var g(t, ω)dW (t, ω) = E ( g(t, ω)dW (t, ω))
0 0

h³ N
X ³ ´´2i
= E lim g(ti−1, ω) W (ti, ω) − W (ti−1, ω)
N →∞
i=1

Stochastic Systems, 2010 13

 Stochastic Integrals

N X
X N
= lim E[g(ti−1, ω)g(tj−1, ω)
N →∞
i=1 j=1

(W (ti, ω) − W (ti−1, ω))(W (tj , ω) − W (tj−1, ω))]

N
X ³ ´2
2
= lim E[g (ti−1, ω)] E[ W (ti, ω) − W (ti−1, ω) ]
N →∞
i=1

N
X 2
= lim E[g (ti−1, ω)] (ti − ti−1)
N →∞
i=1
Z T
2
= E[g (t, ω)]dt . (17)
0

Stochastic Systems, 2010 14

 Stochastic Integrals

The calculation of the variance of the Itô Integrals shows two important properties:

h³ R ´2i RT h 2 i
T
• E 0
g(t, ω)dW (t, ω) = 0 E g (t, ω) dt
RT
• 0
E[g 2(t, ω)]dt < ∞
The second property is the condition of existence for Itô integrals. The next property is
the linearity of Itô integrals:
Z T
[a1 g1(t, ω) + a2 g2(t, ω)]dW (t, ω)
0
Z T Z T
= a1 g1(t, ω)dW (t, ω) + a2 g2(t, ω)dW (t, ω) , (18)
0 0

for numbers a1, a2 and stochastic functions g1(t, ω), g2(t, ω).

Stochastic Systems, 2010 15

 Itô’s lemma

As mentioned shown in the second example, the rules of classical calculus are not valid
for stochastic integrals and differential equations. It is the equivalent to the chain rule
in classical calculus. The problem can be stated as follows:
Given a stochastic differential equation

dX(t) = f (t, X(t))dt + g(t, X(t))dW (t) , (19)

and another process Y (t) which is a function of X(t),

Y (t) = φ(t, X(t)) ,

where the function φ(t, X(t)) is continuously differentiable in t and twice continuously
differentiable in X , find the stochastic differential equation for the process Y (t):

dY (t) = f˜(t, X(t))dt + g̃(t, X(t))dW (t) .

Stochastic Systems, 2010 16

 Itô’s lemma

In the case when we assume that g(t, X(t)) = 0, we know the result: the chain rule
for standard calculus. The result is given by

dy(t) = (φt(t, x) + φx(t, x)f (t, x))dt . (20)

In the case of stochastic problems, we reason as follows: The Taylor expansion of

φ(t, X(t)) yields

1 2
dY (t) = φt(t, X)dt + φtt(t, X)dt + φx(t, X)dX(t)
2
1 2
+ φxx(t, X)(dX(t)) + h.o.t . (21)
2

Stochastic Systems, 2010 17

 Itô’s lemma

We use (19) for dX(t) and get

dY (t) = φt(t, X)dt + φx(t, X)[f (t, X(t))dt + g(t, X(t))dW (t)]
1 ³
2 2 2 2 2
+φtt(t, X)dt + φxx(t, X) f (t, X(t))dt + g (t, X(t))dW (t)
2
´
+2f (t, X(t))g(t, X(t))dt dW (t) + h.o.t . (22)

The differentials of higher order (dt, dW ) become fast zero, dt2 → 0 and
dtdW (t) → 0. The stochastic term dW 2(t) according to the rules of Brownian
motion is given as
2
dW (t, ω) = dt . (23)

Stochastic Systems, 2010 18

 Itô’s lemma

Omitting higher order terms and using the properties of Brownian motion, we arrive at

1 2
dY (t) = [φt(t, X) + φx(t, X)f (t, X(t)) + φxx(t, X)g (t, X(t))]dt
2
+φx(t, X)g(t, X(t))dW (t) . (24)

Reordering the terms yields the scalar version of Itô’s Lemma:

dY (t) = f˜(t, X(t))dt + g̃(t, X(t))dW (t) , (25)

f˜(t, X(t)) = φt(t, X) + φx(t, X)f (t, X(t))
1 2
+ φxx(t, X)g (t, X(t)) , (26)
2
g̃(t, X(t)) = φx(t, X)g(t, X(t)) . (27)

Stochastic Systems, 2010 19

 Itô’s lemma

The term 12 φxx(t, X)g 2(t, X(t)) is often called the Itô corretion term, since this
does not occur in the det. case.
We apply Itôs formula for the following problem: φ(t, X) = X 2 with the SDE
dX(t) = dW (t). From the SDE, we get X(t) = W (t) and calculate the partial
∂φ(t,X) ∂ 2 φ(t,X) ∂φ(t,X)
derivatives of ∂X = 2X , ∂X 2
= 2, and ∂t = 0. The Itô lemma yields

2
d(W (t)) = 1dt + 2W (t)dW (t) . (28)

We rewrite the equation and use W (0) = 0

Z t
2
W (t) = 1t + 2 W (t)dW (t) ,
0
Z t
1 2 1
W (t)dW (t) = W (t) − t . (29)
0 2 2

Stochastic Systems, 2010 20

 Itô’s lemma

We now allow that the process X(t) is in Rn. We let W (t) be an m-dimensional
standard Brownian motion and f (t, X(t)) ∈ Rn and g(t, X(t)) ∈ Rn×m. Consider
a scalar process Y (t) defined by Y (t) = φ(t, X(t)), where φ(t, X) is a scalar
function which is continuously differentiable with respect to t and twice continuously
differentiable with respect to X . The Itô formula can be written in vector notation as
follows:

dY (t) = f˜(t, X(t))dt + g̃(t, X(t))dW (t) , (30)

f˜(t, X(t)) = φt(t, X(t)) + φx(t, X(t)) · f (t, X(t))
1 ³ T
´
+ tr φxx(t, X(t))g(t, X(t))g (t, X(t))) , (31)
2
g̃(t, X(t)) = φx(t, X(t)) · g(t, X(t)) , (32)

where “tr” denotes the trace operator.

Stochastic Systems, 2010 21

 Itô’s lemma

Consider the following stochastic differential equation:

dS(t) = µ S(t)dt + σ S(t)dW (t) , (33)

We want to find the SDE for the process Y related to S as follows: Y (t) = φ(t, S) =
∂φ(t,S) ∂ 2 φ(t,S) ∂φ(t,S)
ln(S(t)) . The partial derivatives are: ∂S = S1 , ∂S 2
= − S12 , and ∂t = 0.
Therefore, according to Itô we get,
³ ∂φ(t, S) ∂φ(t, S) 1 ∂ 2φ(t, S) 2 2 ´
dY (t) = + µS(t) + σ S (t) dt
∂t ∂S 2 ∂S 2
³ ∂φ(t, S) ´
+ σS(t) dW (t) , (34)
∂S
1 2
dY (t) = (µ − σ )dt + σdW (t) . (35)
2

Stochastic Systems, 2010 22

 Itô’s lemma

Since the right hand side of (35) is independent of Y (t), we are able to compute the
stochastic integral:
Z t Z t
1 2
Y (t) = Y0 + (µ − σ )dt + σdW , (36)
0 2 0
1 2
Y (t) = Y0 + (µ − σ )t + σW (t) . (37)
2
Since Y (t) = ln S(t) we have found a solution for S(t) :

1 2
ln(S(t)) = ln(S(0)) + (µ − σ )t + σW (t) , (38)
2
(µ− 1 2
S(t) = S(0)e 2 σ )t+σW (t) , (39)

where W (t) is a standard BM.

Stochastic Systems, 2010 23

 Itô’s lemma

Show for U (t) = X1(t)X2(t) with

dX1(t) = f1(t, X1)dt + g1(t, X1)dW (t) ,

dX2(t) = f2(t, X2)dt + g2(t, X2)dW (t) ,

that following formula is valid:

dU (t) = dX1(t)X2(t) + X1(t)dX2(t) + g1(t, X1)g2(t, X2)dt (40)

We show that we obtain the same result as in the previous formula by apply Itô’s
lemma. By (40) liefert

dU (t) = [X2(t)f1(t, X1) + X1(t)f2(t, X2) + g1(t, X1)g2(t, X2)]dt

+[X2(t)g1(t, X1) + X1(t)g2(t, X2)]dW (t)

Stochastic Systems, 2010 24

 Itô’s lemma

· ¸
∂ 2U 0 1
The partial derivatives of U are : ∂U
∂X = (X2(t), X1(t))T , ∂X 2
= and
1 0
∂U
∂t = 0.

∂U ∂U T
dU (t) = [ + [f1(t, X1), f2(t, X2)]
∂t ∂X
· ¸´
1 ³ ∂ 2U g1(t, X1)2 g1(t, X1)g2(t, X2)
+ tr ]dt
2 ∂X 2 g1(t, X1)g2(t, X2) g2(t, X2)2
∂U T
+ [g1(t, X1), g2(t, X2)] dW (t)
∂X
= [X2(t)f1(t, X1) + X1(t)f2(t, X2) + g1(t, X1)g2(t, X2)]dt
+[X2(t)g1(t, X1) + X1(t)g2(t, X2)]dW (t)

Stochastic Systems, 2010 25

 Stochastic Differential Equations (SDE)

We classify SDEs into two large groups, linear SDEs and non-linear SDEs. Furthermore,
we distinguish between scalar linear and vector-valued linear SDEs.
We start with the easy case, the scalar linear linear SDEs. An SDE

dX(t) = f (t, X(t))dt + g(t, X(t))dW (t) , (41)

for a one-dimensional stochastic process X(t) is called a linear (scalar) SDE if and
only if the functions f (t, X(t)) and g(t, X(t)) are affine functions of X(t) ∈ R and
thus

f (t, X(t)) = A(t)X(t) + a(t) ,

g(t, X(t)) = [B1(t)X(t) + b1(t), · · · , Bm(t)X(t) + bm(t)] ,

where A(t), a(t) ∈ R, W (t) ∈ Rm is an m-dimensional Brownian motion, and

Bi(t), bi(t) ∈ R, i = 1, · · · , m. Hence, f (t, X(t)) ∈ R and g(t, X(t)) ∈ R1×m.

Stochastic Systems, 2010 26

 Stochastic Differential Equations (SDE)

The linear SDE possesses the following solution

³ Z t h m
X i
−1
X(t) = Φ(t) x0 + Φ (s) a(s) − Bi(s)bi(s) ds
0 i=1
m Z
X t ´
−1
+ Φ (s)bi(s)dWi(s) , (42)
i=1 0

where we denote Φ(t) as the fundamental matrix, which we obtain from

³Z t h Xm
Bi2(s) i Xm Z t ´
Φ(t) = exp A(s) − ds + Bi(s)dWi(s) , (43)
0 i=1
2 i=1 0

The solution is similar to the solution of ODEs.

Stochastic Systems, 2010 27

 Stochastic Differential Equations (SDE)

Let us assume that W (t) ∈ R, a(t) = 0, b(t) = 0, A(t) = A, B(t) = B . We

want to compute the solution of the SDE

dX(t) = AX(t)dt + BX(t)dW (t) , X(t) = x0 , (44)

We can solve it using (42) and (43):

1 B 2 )t+BW (t)
(A− 2
Φ(t) = e , (45)

and (42) is easy to calculate since

1 B 2 )t+BW (t)
(A− 2
x(t) = Φ(t)x0 = x0e . (46)

Stochastic Systems, 2010 28

 Stochastic Differential Equations (SDE)

The expectation m(t) = E[X(t)]and the second moment P (t) = E[X 2(t)] for
m
X
dX(t) = (A(t)X(t) + a(t))dt + (Bi(t)X(t) + b(t))dWi(t) . (47)
i=1

can be calculated by solving the following system of ODEs:

ṁ(t) = A(t)m(t) + a(t) , m(0) = x0 , (48)

³ Xm ´ ³ m
X ´
2
Ṗ (t) = 2A(t) + Bi (t) P (t) + 2m(t) a(t) + Bi(t)bi(t)
i=1 i=1
m
X ´
2 2
+ bi (t) , P (0) = x0 . (49)
i=1

Stochastic Systems, 2010 29

 Stochastic Differential Equations (SDE)

The ODE for the expectation is derived by applying the expectation operator on both
sides of (42).
m
X
E[dX(t)] = E[(A(t)X(t) + a(t))dt + (Bi(t)X(t) + bi(t))dWi(t) ]
i=1

E[dX(t)] = (A(t) E[X(t)] +a(t))dt

| {z } | {z }
dm(t) =m(t)

m
X
+ E[(Bi(t)X(t) + bi(t))] E[dWi(t) ]
| {z }
i=1 =0

dm(t) = (A(t)m(t) + a(t))dt . (50)

Stochastic Systems, 2010 30

 Stochastic Differential Equations (SDE)

In order to compute the second moment, we need to derive the SDE for Y (t) = X 2(t):

h m ³
X ´2i
dY (t) = 2X(t)(A(t)X(t) + a(t)) + Bi(t)X(t) + bi(t) dt
i=1
m ³
X ´
+2X(t) Bi(t)X(t) + bi(t) dWi(t) (51)
i=1
h m ³
X
2 2 2
dY (t) = 2A(t)X (t) + 2X(t)a(t) + Bi (t)X (t) + 2Bi(t)bi(t)X(t)
i=1
´i m ³
X ´
2
+bi (t) dt + 2X(t) Bi(t)X(t) + bi(t) dWi(t) (52)
i=1

Stochastic Systems, 2010 31

 Stochastic Differential Equations (SDE)

Furthermore, we apply the expectation operator to (52) and use P (t) = E[X 2(t)] =
E[Y (t)] and m(t) = E[X(t)].

h m ³
X
2 2 2
E[dY (t)] = 2A(t)E[X (t)] + 2a(t)E[X(t)] + Bi (t)E[X (t)]
i=1
´i
2
+2Bi(t)bi(t)E[X(t)] + bi (t) dt
h m ³
X ´ i
+E 2X(t) Bi(t)X(t) + bi(t) dWi(t)
i=1
h
dP (t) = 2A(t)P (t) + 2a(t)m(t)
m ³
X ´i
2 2
+ Bi (t)P (t) + 2Bi(t)bi(t)m(t) + bi (t) dt
i=1

Stochastic Systems, 2010 32

 Stochastic Differential Equations (SDE)

In the case that Bi(t) = 0, i = 1, . . . , m, we are able to directly compute the

distribution. The scalar linear SDE
m
X
dX(t) = (A(t)X(t) + a(t))dt + bi(t)dWi(t), (53)
i=1

with X(0) = x0 is normaly distributed

P (X(t)|x0) ∼ N (m(t), V (t)) with expected value m(t) and variance V (t), which
are solutions of the following ODEs,

ṁ(t) = A(t)m(t) + a(t) , m(0) = x0 , (54)

m
X 2
V̇ (t) = 2A(t)V (t) + bi (t) , V (0) = 0 . (55)
i=1

Stochastic Systems, 2010 33

 Stochastic Differential Equations (SDE)

There are some specific scalar linear SDEs which are found to be quite useful in practice.
The simplest case of SDE is where the drift and the diffusion coefficients are independent
of the information received over time

dS(t) = µdt + σdW (t) , S(0) = S0 . (56)

This model has been used to simulate commodity prices, such as metals or agricultural
products.
The mean is E[S(t)] = µt + S0 and the variance Var[S(t)] = σ 2t. S(t) possesses
a behavior of fluctuations around the straight line S0 + µt.The process is normally
distributed with the given mean and variance.

Stochastic Systems, 2010 34

 Stochastic Differential Equations (SDE)

The standard model of stock prices is the geometric Brownian motion as given by

dS(t) = µS(t)dt + σS(t)dW (t, ω) , S(0) = S0 .

2
The mean is given by E[S(t)] = S0eµt and its variance by Var[S(t)] = S02e2µt(eσ t −
1). This model forms the starting point for the famous Black-Scholes formula for option
pricing. The geometric Brownian motion has two main features which make it popular
for stock
The first property is that S(t) > 0 for all t ∈ [0, T ] and the second is that all returns
are in scale with the current price. This process has a log-normal probability density
function.

Stochastic Systems, 2010 35

 Stochastic Differential Equations (SDE)

Another very popular class of SDEs are mean reverting linear SDEs. The model is
obtained by

dS(t) = κ[µ − S(t)]dt + σ dW (t, ω) , S(0) = S0 . (57)

A special case of this SDE where µ = 0 is called Ohrnstein-Uhlenbeck process.

Equation (57) models a process which naturally falls back to its equilibrium level of µ.
The expected price is E[S(t)] = µ − (µ − S0)e−κ t and the variance is

σ2 ³ −2κ t
´
Var[S(t)] = 1−e .
2κ

Stochastic Systems, 2010 36

 Stochastic Differential Equations (SDE)

In the long run, the following (unconditional) approximations are valid

lim E[S(t)] = µ
t→∞

and
σ2
lim Var[S(t)] = .
t→∞ 2κ
2
This analysis shows that the process fluctuates around µ and has a variance of σ2κ
which depends on the parameter κ: the higher κ, the lower the variance.
This is obvious since the higher κ, the faster the process reverts back to its mean
value.

This process is a stationary process which is normally distributed.

Stochastic Systems, 2010 37

 Stochastic Differential Equations (SDE)

A popular extension is where the diffusion term is in scale with the current value, i.e.,
the geometric mean reverting process:

dS(t) = κ[µ − S(t)]dt + σS(t)dW (t, ω) , S(0) = S0 .

In this model S(t) ≥ 0, if S0 ≥ 0, µ > 0, and κ > 0.

The first mean reversion model(57) may produce negative values even for µ > 0.

Since the second mean-reversion model has always positive realizations, it is also
called log-normal mean reversion. This type of model is used to model interest rate or
volatilities.

Stochastic Systems, 2010 38

 Stochastic Differential Equations (SDE)

In control engineering science, the most important (scalar) case is

m
X
dX(t) = (A(t)X(t) + C(t)u(t)) dt + bi(t) dWi . (58)
i=1

In this equation, X(t) is normally distributed because the Brownian motion is just
multiplied by time-dependent factors.

When we compute an optimal control law for this SDE, the deterministic optimal control
law (ignoring the Brownian motion) and the stochastic optimal control law are the same.

This feature is called certainty equivalence. For this reason, the stochastics are often
ignored in control engineering.

Stochastic Systems, 2010 39

 Stochastic Differential Equations (SDE)

The logical extension of scalar SDEs is to allow X(t) ∈ Rn to be a vector. The rest of
this section proceeds in a similar fashion as for scalar linear SDEs. A stochastic vector
differential equation

dX(t) = f (t, X(t))dt + g(t, X(t))dW (t)

with the initial condition X(0) = x0 ∈ Rn for an n-dimensional stochastic process

X(t) is called a linear SDE if the functions f (t, X(t)) ∈ Rn and g(t, X(t)) ∈ Rn×m
are affine functions of X(t) and thus

f (t, X(t)) = A(t)X(t) + a(t) ,

g(t, X(t)) = [B1(t)X(t) + b1(t), · · · , Bm(t)X(t) + bm(t)] ,

where A(t) ∈ Rn×n, a(t) ∈ Rn, W (t) ∈ Rm is an m-dimensional Brownian motion,

and Bi(t) ∈ Rn×n, bi(t) ∈ Rn.

Stochastic Systems, 2010 40

 Stochastic Differential Equations (SDE)

Alternatively, the vector-valued linear SDE can be written as

m
X
dX(t) = (A(t)X(t) + a(t))dt + (Bi(t)X(t) + bi(t))dWi(t) . (59)
i=1

A common extension of the above equation is the following form of a controlled

stochastic differential equation as given by

dX(t) = (A(t)X(t) + C(t)u(t) + a(t)) dt

m
X
+ (Bi(t)X(t) + Di(t)u(t) + bi(t)) dWi , (60)
i=1

where u(t) ∈ Rk , C(t) ∈ Rn×k , Di(t) ∈ Rn×k .

Stochastic Systems, 2010 41

 Stochastic Differential Equations (SDE)

The linear SDE (59) has the following solution:

³ Z t h m
X i
−1
X(t) = Φ(t) x0 + Φ (s) a(s) − Bi(s)bi(s) ds
0 i=1
m Z
X t ´
−1
+ Φ (s)bi(s)dWi(s) , (61)
i=1 0

where the fundamental matrix Φ(t) ∈ Rn×n is the solution of the homogenous
stochastic differential equation.

Stochastic Systems, 2010 42

 Stochastic Differential Equations (SDE)

The fundamental matrix Φ(t) ∈ Rn×n is the solution of the homogenous stochastic
differential equation:
m
X
dΦ(t) = A(t)Φ(t)dt + Bi(t)Φ(t)dWi(t) , (62)
i=1

with initial condition Φ(0) = I , I ∈ Rn×n e now prove that (61) and (62) are
solutions of (59). We rewrite (61) as
³ Z t ´
−1
X(t) = Φ(t) x0 + Φ (t)dY (t)
0
h m
X i m
X
dY (t) = a(t) − Bi(t)bi(t) dt + bi(t)dWi(t) .
i=1 i=1

Stochastic Systems, 2010 43

 Stochastic Differential Equations (SDE)

³ Z t ´
−1
X(t) = Φ(t)Z(t) , Z(t) = x0 + Φ (t)dY (t)
0
−1
dZ(t) = Φ (t)dY (t)

We use the Itô formula to calculate X(t) = Φ(t)Z(t):

m
X −1
dX(t) = Φ(t)dZ(t) + dΦ(t)Z(t) + Bi(t)Φ(t)Φ(t) bi(t)dt
i=1
m
X m
X
= dY (t) + A(t)Φ(t)Z(t)dt + Bi(t)Φ(t)Z(t)dWi(t) + Bi(t)bi(t)dt
i=1 i=1

Stochastic Systems, 2010 44

 Stochastic Differential Equations (SDE)

Noting that Z(t) = Φ−1(t)X(t) and using the SDE for Y (t), we get
m
X m
X
dX(t) = dY (t) + A(t)Φ(t)Z(t)dt + Bi(t)Φ(t)Z(t)dWi(t) + Bi(t)bi(t)dt
i=1 i=1
h m
X i m
X
= a(t) − Bi(t)bi(t) dt + bi(t)dWi(t) + A(t)X(t)dt
i=1 i=1
m
X m
X
+ Bi(t)X(t)dWi(t) + Bi(t)bi(t)dt
i=1 i=1
m
X
= [a(t) + A(t)X(t)]dt + (Bi(t)X(t) + bi(t))dWi(t) .
i=1

This completes the proof.

Stochastic Systems, 2010 45

 Stochastic Differential Equations (SDE)

The expectation m(t) = E[X(t)] ∈ Rn and the second moment matrix P (t) =
E[X(t)X T (t)] ∈ Rn×n can be computed as follows:

ṁ(t) = A(t)m(t) + a(t) , m(0) = x0 , (63)

T T T
Ṗ (t) = A(t)P (t) + P (t)A (t) + a(t)m (t) + m(t)a (t)
m
X T T
+ [Bi(t)P (t)Bi (t) + Bi(t)m(t)bi (t)
i=1
T T T T
+bi(t)m (t)Bi (t) + bi(t)bi(t) ] , P (0) = x0x0 . (64)

The covariance matrix for the system of linear SDEs is given by als
T
V (t) = Var{x(t)} = P (t) − m(t)m (t) . (65)

Stochastic Systems, 2010 46

 Stochastic Differential Equations (SDE)

The special case

m
X
dX(t) = (A(t)X(t) + a(t))dt + bi(t)dWi(t)
i=1

with the initial condition X(0) = x0 ∈ Rn is normally distributed, i.e.,

P (X(t)|x0) ∼ N (m(t), V (t))

where

ṁ(t) = A(t)m(t) + a(t) m(0) = x0

m
X
T T
V̇ (t) = A(t)V (t) + V (t)A (t) + bibi (t) V (0) = 0 .
i=1

Stochastic Systems, 2010 47

 Stochastic Differential Equations (SDE)

As first example of a linear vector valued SDE, we consider a two dimensional geometric
Brownian motion:
³ ´
dS1(t) = µ1S1(t)dt + S1(t) σ11dW1(t) + σ12dW2(t) , (66)
³ ´
dS2(t) = µ2S2(t)dt + S2(t) σ21dW1(t) + σ22dW2(t) . (67)

Written in matrix form S = (S1, S2)T , the same SDE is given as:
µ ¶ µ ¶ µ ¶ µ ¶
µ1 0 0 σ11 0 σ12 0
A(t) = a(t) = B1(t) = B2(t) =
0 µ1 0 0 σ21 0 σ22

Both processes S1(t) and S2(t) are correlated if σ12 = σ21 6= 0. This model can be
easily extended to n processes.

Stochastic Systems, 2010 48

 Stochastic Differential Equations (SDE)

The observed volatility for real existing price processes, such as stocks or bonds is itself
a stochastic process. The following model describes this observation:

dP (t) = µdt + σ(t)dW1(t) , P (0) = P0 ,

dσ(t) = κ(θ − σ(t))dt + σ(t)σ1dW2(t) , σ(0) = σ0 .

where θ is the average volatility, σ1 a volatility, and κ the mean reversion rate of
the volatility process σ(t). If this model is used for stock prices, the transformation
P (t) = ln(S(t)) is useful. The two Brownian motions dW1(t) and dW2(t) are
correlated, hence corr[dW1(t), dW2(t)] = ρ. This model captures the behavior of
real existing prices better and its distribution of returns shows “fatter tails”.

Stochastic Systems, 2010 49

 Stochastic Differential Equations (SDE)

Die system (68) can be rewritten as linear SDE:

µ ¶ µ ¶ µ ¶
0 0 µ 0 1
A(t) = a(t) = B1(t) =
0 −κ κθ 0 σ1 ρ
µ ¶
0 p0
B2(t) = ,
0 σ 1 1 − ρ2

wobei x(t) = (P (t), σ(t))T . The system (68) has the property, that the variance
of P (t) depends on the initial condition σ0 For the parameters µ = 0.1, κ = 2,
θ = 0.2, σ1 = 0.5 and ρ = 0.5, we calculate the standard deviation of P (t) with
σ0 = 0.1 and alternatively with σ0 = 0.8. The expected value of σ(t) has the
following evaluation over time m(t) = θ + (σ0 − θ)e−κt and thus the variance of
P (t) depends on σ0.

Stochastic Systems, 2010 50

 Stochastic Differential Equations (SDE)

0.7
σ0=0.1
σ0=0.8
0.6

0.5
Standardabweichung

0.4

0.3

0.2

0.1

0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
time

Abbildung 1: Stand. dev. of P (t) for different initial conditions of σ(t)

Stochastic Systems, 2010 51

 Stochastic Differential Equations (SDE)

In comparison with linear SDEs, nonlinear SDEs are less well understood. No general
solution theory exists. And there are no explicit formulae for calculating the moments.
In this section, we show some examples of nonlinear SDEs and their properties.
In general, a scalar square root process can be written as

dX(t) = f (t, X(t))dt + g(t, X(t))dW (t)

with
f (t, X(t)) = A(t)X(t) + a(t)
q
g(t, X(t)) = B(t) X(t) ,

where A(t), a(t), and B(t) are real scalars. The nonlinear mean reverting SDEs differ
from the linear scalar equations by their nonlinear diffusion term. For this process, the
distribution and moments can be calculated.

Stochastic Systems, 2010 52

 Stochastic Differential Equations (SDE)

For a specific square root process with A(t) = 0, a(t) = 1 and B(t) = 2 we are
able to derive the analytical solution: The SDE
q
dX(t) = 1dt + 2 X(t)dW (t) , X(0) = xo ,

has the solution X(t) = (W (t) + x0)2We verify the solution using Itô formula. We
use Φ(t) = X(t) = (Y (t) + x0)2 and dY (t) = dW (t). The partial derivatives are
Φt = 0, ΦY = 2(Y (t) + x0), and ΦY Y = 2. Thus

1
dΦ(t) = [Φt + ΦY · 0 + ΦY Y · 1]dt + ΦY · 1dW (t) ,
2
q
dΦ(t) = 1dt + 2(Y (t) + x0)dW (t) , ⇒ dX(t) = 1dt + 2 X(t)dW (t) ,
p
since X(t) = Y (t) + x0.

Stochastic Systems, 2010 53

 Stochastic Differential Equations (SDE)

Another widely used mean reversion model is obtained by

q
dS(t) = κ[µ − S(t)]dt + σ S(t)dW (t) , S(0) = S0 . (68)

This model is also known as the Cox-Ross-Ingersol processes.The process shows a

less volatile behavior than its linear geometric counterpart and it has a non-central
chi-square distribution. The process is often used to model short-term interest rates or
stochastic volatility processes for stock prices. Another often used square root process
is similar to the geometric Brownian motion, but with a square root diffusion term
instead of the linear diffusion term. Its model is given by
q
dS(t) = µS(t)dt + σ S(t)dW (t) , S(0) = S0 . (69)

Stochastic Systems, 2010 54

 Stochastic Differential Equations (SDE)

µt
The expected value
³ for (69) is
´ E [S(t)] = S 0 e and the variance is obtained by
2
σ S
Var[S(t)] = µ 0 e2µt − eµt .
Another widely used mean reversion model is obtained by

dS(t) = κS(t)[µ − ln(S(t))]dt + S(t)σdW (t) . (70)

Using the transformation P (t) = ln(S(t)) yields the linear mean reverting and
normally distributed process P (t):

σ2
dP (t) = κ[(µ − ) − P (t)]dt + σdW (t) , (71)
2κ
Because of the transformation, S(t) is log-normally distributed. This model is used
to model stock prices, stochastic volatilities, and electricity prices. Because S(t) is
log-normally distributed, S(t) is always positive.

Stochastic Systems, 2010 55

 Stochastic Differential Equations (SDE)

In this part, we introduce three major methods to compute solution of SDEs.

• The first method is based on the Itô integral and has already been used for linear
solutions.
• We introduce numerical methods to compute path-wise solutions of SDEs.
• The third method is based on partial differential equations, where the problem of
finding the probability density function of the solution is transformed into solving a
partial differential equation.

Stochastic Systems, 2010 56

 Stochastic Differential Equations (SDE)

The stochastic process X(t) governed by the stochastic differential equation

dX(t) = f (t, X(t))dt + g(t, X(t))dW (t)

X(0) = X0

is explicitly described by the integral form

Z t Z t
X(t, ω) = X0 + f (s, X(s)) ds + g(s, X(s)) dW (s) ,
0 0

where the first integral is a path-wise Riemann integral and the second integral is an
Itô integral.
In this definition, it is assumed that the functions f (t, X(t)) and g(t, X(t)) are
sufficiently smooth in order to guarantee the existence of the solution X(t).

Stochastic Systems, 2010 57

 Stochastic Differential Equations (SDE)

There are several ways of finding analytical solutions. One way is to guess a soluti-
on and use the Itô calculus to verify that it is a solution for the SDE under consideration.

We assume that the following nonlinear SDE

q
dX(t) = dt + 2 X(t) dW (t) ,

has the solution

p 2
X(t) = (W (t) + X0) .

In order to verify this√claim, we use the Itô calculus. We have

√ X(t) = φ(W ) where
φ(W ) = (W (t) + X0)2, so that φ0(W ) = 2(W (t) + X0) and φ00(W ) = 2.

Stochastic Systems, 2010 58

 Stochastic Differential Equations (SDE)

Using Itô’s rule, we get

dX(t) = fe(t, X)dt + g

e(t, X)dW (t)
1 00 q
fe(t, X)
0 2
= φ(W ) 1 + φ (W )(2 X(t)) = 1
2
q p
0
g
e(T, X) = φ (W )(2 X(t)) = 2(W (t) + X0) .
√ 2 √ p
Since X(t) = (W (t) + X0) we know that (W (t) + X0) = X(t) and thus
the Itô calculation generated the original SDE where we started at.

Stochastic Systems, 2010 59

 Stochastic Differential Equations (SDE)

For some classes of SDEs, analytical formulas exist to find the solution, e.g. consider
the following SDE:

dX(t) = f (t, X(t))dt + σ(t)dW (t) , X(0) = x0 (72)

where X(t) ∈ Rn, f (t, X(t)) ∈ Rn is an arbitrary function, σ(t) ∈ Rn×m and
dW (t) ∈ Rm. This class of SDEs has the following general solution:

X(t) = Y (t) + F (t) (73)

dY (t) = f (t, Y (t) + F (t))dt , Y (0) = x0 (74)
dF (t) = σ(t)dW (t) , F (0) = 0 . (75)
Rt
The SDE for F (t) can be integrated, i.e. F (t) = 0
σ(s)dW (s). When σ(t) = σ
than F (t) = σW (t).

Stochastic Systems, 2010 60

 Stochastic Differential Equations (SDE)

SinceF (t) is know,, we are able to solve for Y (t) in in function of F (t).
Using Itô lemman, we show that X(t) = Y (t) + F (t) and this solves the SDE

dX(t) = dY (t) + dF (t) = f (t, Y (t) + F (t))dt + σ(t)dW (t)

= f (t, X(t))dt + σ(t)dW (t) (76)

This solution is not very suprising, since X(t) is the sum of the process of Y (t) and
the BM of F (t).

Stochastic Systems, 2010 61

 Stochastic Differential Equations (SDE)

For another class of SDEs, exist an analytical formula for their solution:

dX(t) = f (t, X(t))dt + c(t)X(t)dW (t) , X(0) = x0 , (77)

where f (t, X(t)) ∈ R, c(t) ∈ R and dW ∈ R. DThe solution can be derived as

follows:
−1
X(t) = F (t)Y (t) (78)
2
dF (t) = F (t)c (t)dt − F (t)c(t)dW (t) , F (0) = 1 (79)
−1
dY (t) = F (t)f (t, F Y (t))dt (80)

The proof is similar to the first case, sice the diffusion is linear.

Stochastic Systems, 2010 62

 Stochastic Differential Equations (SDE)

Calculate the analytical solution for

dt
dX(t) = + αX(t)dW (t) , X(0) = x0 .
X(t)
1 α2 t−αW (t) F (t) F 2(t)
F (t) = e2 , dY (t) = −1 dt = dt
F (t)Y Y
Z t
2 1 2 2
dY (t)Y (t) = F (t)dt , Y (t) = F (s)ds + C0
2 0
³ Z t ´1
2 α2 s−2αW (s) 2
Y (t) = x0 + 2 e ds
0
Z
−1 α 2 t+αW (t) ³ 2 t
α2 s−2αW (s)
´1
2
X(t) = e 2 x0 +2 e ds
0

Stochastic Systems, 2010 63

 Stochastic Differential Equations (SDE)

However, most SDEs, especially nonlinear SDEs, do not have analytical solutions so
that one has to resort to numerical approximation schemes in order to simulate sample
paths of solutions to the given equation.
The simplest scheme is obtained by using a first-order approximation. This is called the
Euler scheme

X(tk ) = X(tk−1) + f (tk−1, X(tk−1))∆t + g(tk−1, X(tk−1))∆W (tk ) .

The Brownian motion term can be approximated as follows:

√
∆W (tk ) = ²(tk ) ∆t ,

where the ²(.) is a discrete-time Gaussian white process with mean 0 and standard
deviation 1.

Stochastic Systems, 2010 64