I am working with two correlated SDE's and I was wondering if I could use two different discretization schemes for them. Is there maybe a reference of this being done? And can something be said about the convergence order? Is it better to use two Euler discretizations in stead of 1 Euler and 1 Milstein?

  • $\begingroup$ I think your bias is reduced for every dimension in which you are using Milstein. But, if the two processes have "canceling" effects on the final result, you might be better off matching their discretization even if the match must be done at Euler level. $\endgroup$ – Brian B Aug 10 '14 at 19:45

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.