English

Speeding up the Euler scheme for killed diffusions

Numerical Analysis 2021-10-01 v2 Numerical Analysis Probability

Abstract

Let XX be a linear diffusion taking values in (,r)(\ell,r) and consider the standard Euler scheme to compute an approximation to E[g(XT)1[T<ζ]]\mathbb{E}[g(X_T)\mathbf{1}_{[T<\zeta]}] for a given function gg and a deterministic TT, where ζ=inf{t0:Xt(,r)}\zeta=\inf\{t\geq 0: X_t \notin (\ell,r)\}. It is well-known since \cite{GobetKilled} that the presence of killing introduces a loss of accuracy and reduces the weak convergence rate to 1/N1/\sqrt{N} with NN being the number of discretisatons. We introduce a drift-implicit Euler method to bring the convergence rate back to 1/N1/N, i.e. the optimal rate in the absence of killing, using the theory of recurrent transformations developed in \cite{rectr}. Although the current setup assumes a one-dimensional setting, multidimensional extension is within reach as soon as a systematic treatment of recurrent transformations is available in higher dimensions.

Keywords

Cite

@article{arxiv.2107.03534,
  title  = {Speeding up the Euler scheme for killed diffusions},
  author = {Umut Çetin and Julien Hok},
  journal= {arXiv preprint arXiv:2107.03534},
  year   = {2021}
}

Comments

Some typos and errors in the earlier version are corrected