An application of the multiplicative Sewing Lemma to the high order weak approximation of stochastic differential equations (Q6080380)

The authors derive an arbitrary high order weak approximation result for stochastic differential equations as an application of a version of the multiplicative sewing lemma. In contrast to the previous results in this direction, this lemma does not use any reflexivity assumption. This allows to deduce weak convergence with rather weak assumptions in terms of the expected signature of the approximating driving signal. The new approximation covers, in particular, the cubature approximations on Wiener space. The main assumption on the noise approximations (it is Assumption 2.3 in the paper) is formulated in a comparatively complicated form, however, the authors provide a sufficient number of convincing examples.