I'm looking for a result of the form: Let $B_\epsilon$ denote a "natural" smooth $\epsilon$-approximation to an $n$-dimensional Brownian motion $B$ (e.g. by mollification or simply piecewise linear) with $\sigma$ and $\mu$ smooth coefficients, and denote by $X_\epsilon$ the solution to the random ODE $$dX_\epsilon = \mu(X_\epsilon)dt + \sigma(X_\epsilon)dB_\epsilon.$$

Then $X_\epsilon$ converges in probability, as $\epsilon \to 0$, so the solution of the SDE $$dX=\mu(X)dt+\sigma(X)\circ dB.$$

In particular I'm seeking results with no ellipticity requirement on $\sigma$. Many thanks.


Have you seen the origianl Wong-Zakai paper?


  • $\begingroup$ Can you give more details about the paper and the result that might interest the OP? An answer should remain meaningful even if all the links die. $\endgroup$ – Joonas Ilmavirta Nov 9 '15 at 12:30
  • 2
    $\begingroup$ The original Wong-Zakai paper actually doesn't treat the $n$-dimensional case. In the one-dimensional case, any (continuous) approximation converges, which is not the case in higher dimensions. To the best of my knowledge the first articles treating the $n$-dimensional case are Strook & Varadhan's support theorem paper. (Coincidentally, there's also a paper by McShane with a similar result that appeared in the same proceedings volume.) $\endgroup$ – Martin Hairer Feb 7 '16 at 14:07

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.