w0 =
X0
wi+1 =
wi +
a∆ti +
b∆Wi +
1 bbx(
∆W 2 − ∆ti)
2
i
+
axb∆Zi +
1 (
aax +
1 axxb2)
∆t2
2 2
2
+ (
abx +
1 bxxb2)(
∆Wi∆ti − ∆Zi) (18) where
∆Wi is chosen from
√∆tiN (0
, 1) and
∆Zi is distributed as in the above Strong Order 1.5 Method.
A second approach is to mimic the idea of Runge-Kutta solvers for ordinary differential equations. These solvers replace the explicit higher derivatives in the Ito-Taylor solvers with extra function evaluations at interior points of the current solution interval. Platen (1987) proposed the following weak order 2 solver of Runge-Kutta type:
u+ =
wi +
a∆ti +
b ∆ti u− =
wi +
a∆ti − b ∆ti.
0
10
−1
10
error
−2
10
−3
10
−4
10 −2 −1 0
10 10 10
time step t
Fig. 4. The mean error of the estimation of E(
X(
T ))
for SDE (15). The plot compares the Euler-Maruyama method (circles) which has weak order 1, and the weak order 2 Runge-Kutta type method (squares) given in (19). Parameter used were
X(0) = 10
, T = 1
, µ =
−3
, σ = 0
.2.
Fig. 4 compares
the Euler-Maruyama method, which converges with order 1 in the weak sense, to the Weak Order 2 Runge-Kutta-Type Method. Note the difference between strong and weak convergence. In the previous Fig. 3, which
considers strong convergence, the mean error of the estimate of a point
X(
T ) on the solution curve was plotted. In Fig. 4,
on the other hand, the mean error of the estimate of the expected value
E[
X(
T )] is plotted, since we are comparing weak convergence of the methods. The weak orders are clearly revealed by the log-log plot.