The Milstein method of stochastic analysis describes a method for the numerical solution of stochastic differential equations (SDGL), named after the Russian mathematician Grigori Noichowitsch Milstein ( Gorky State University of the Urals ).
algorithm
Look at the Itō SDGL
d
X
t
=
a
(
X
t
)
d
t
+
b
(
X
t
)
d
W.
t
,
{\ displaystyle \ mathrm {d} X_ {t} = a (X_ {t}) \, \ mathrm {d} t + b (X_ {t}) \, \ mathrm {d} W_ {t},}
with initial condition , denoting the Wiener process . If a solution is to be found on the interval , the Milstein method gives an approximation for the true solution on an equidistant grid:
X
0
=
x
0
{\ displaystyle X_ {0} = x_ {0}}
W.
t
{\ displaystyle W_ {t}}
[
0
,
T
]
{\ displaystyle [0, T]}
Y
{\ displaystyle Y}
X
{\ displaystyle X}
Break down the interval into equally long sub-intervals of length :
[
0
,
T
]
{\ displaystyle [0, T]}
N
{\ displaystyle N}
δ
>
0
{\ displaystyle \ delta> 0}
0
=
τ
0
<
τ
1
<
⋯
<
τ
N
=
T
{\ displaystyle 0 = \ tau _ {0} <\ tau _ {1} <\ dots <\ tau _ {N} = T}
and .
δ
=
T
N
{\ displaystyle \ delta = {\ tfrac {T} {N}}}
Set .
Y
0
: =
x
0
{\ displaystyle Y_ {0}: = x_ {0}}
Define for through
Y
n
+
1
{\ displaystyle Y_ {n + 1}}
0
≤
n
<
N
{\ displaystyle 0 \ leq n <N}
Y
n
+
1
: =
Y
n
+
a
(
Y
n
)
δ
+
b
(
Y
n
)
Δ
W.
n
+
1
2
b
(
Y
n
)
b
′
(
Y
n
)
(
(
Δ
W.
n
)
2
-
δ
)
,
{\ displaystyle Y_ {n + 1}: = Y_ {n} + a (Y_ {n}) \ delta + b (Y_ {n}) \ Delta W_ {n} + {\ frac {1} {2}} b (Y_ {n}) b '(Y_ {n}) \ left ((\ Delta W_ {n}) ^ {2} - \ delta \ right),}
in which
Δ
W.
n
=
W.
τ
n
+
1
-
W.
τ
n
{\ displaystyle \ Delta W_ {n} = W _ {\ tau _ {n + 1}} - W _ {\ tau _ {n}}}
and the derivation of relative is. Note that the random variables are independently normally distributed with expectation value 0 and variance .
b
′
{\ displaystyle b '}
b
(
x
)
{\ displaystyle b (x)}
x
{\ displaystyle x}
Δ
W.
n
{\ displaystyle \ Delta W_ {n}}
δ
{\ displaystyle \ delta}
convergence
With the above notation applies to and all , which is why one speaks of first- order convergence . is a Landau symbol .
E.
[
|
Y
n
-
X
(
τ
n
)
|
]
=
O
(
δ
)
{\ displaystyle E [| Y_ {n} -X (\ tau _ {n}) |] = {\ hbox {o}} (\ delta) \;}
δ
→
0
{\ displaystyle \ delta \ to 0}
n
=
0
,
.
.
.
,
N
{\ displaystyle n = 0, ..., N}
O
{\ displaystyle {\ hbox {o}}}
See also
credentials
<img src="https://de.wikipedia.org/wiki/Special:CentralAutoLogin/start?type=1x1" alt="" title="" width="1" height="1" style="border: none; position: absolute;">