f*******n
发帖数: 70 | 1 Right. More precisely, discretize the SDE as:
S(t)-S(t-1)=r*S(t-1)*dt+v*S(t-1)*sqrt(dt)*Normal(0,1),
simulate the N(0,1) r.v. and proceed step by step.
Actually what I usually do is to simulate log(St) first by:
d(lnS)=(r-v^2/2)dt+vdW, and then take the exp, a bit easier to implement.
If W1, W2 are not independent, you need to simulate two correlated N(0,1) r.v.
Oops, it seems to me the "v" in the first equation is the same as the "vt" in
the 2nd(volatility?), then you need to simulate vt first, |
|
