2024/09/05 by Ruishu Liu, Andreas Neuenkirch, Xiaojie Wang · 2 citations
Engineering · Computer Science · #Advanced Numerical Methods in Computational Mathematics #Numerical methods in engineering #Advanced Mathematical Modeling in Engineering
paper · doi:10.1090/mcom/4014
In this paper, we study the strong approximation of scalar stochastic differential equations (SDEs), which take values in a domain and have non-Lipschitz coefficients. By combining a Lamperti-type transformation with a semi-implicit discretization approach and a taming strategy, we construct a domain-preserving scheme that strongly converges under weak assumptions. Moreover, we show that this scheme has strong convergence order <inline-formula content-type="math/mathml"> <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" alttext="1.5"> <mml:semantics> <mml:mn>1.5</mml:mn> <mml:annotation encoding="application/x-tex">1.5</mml:annotation> </mml:semantics> </mml:math> </inline-formula> under additional assumptions on the coefficients of the SDE. In our scheme, the domain preservation is a consequence of the semi-implicit discretization approach, while the taming strategy allows controlling terms of the scheme that admit singularities but are required to obtain the desired order. Our general convergence results are applied to various SDEs from applications, with sub-linearly or super-linearly growing and non-globally Lipschitz coefficients. Numerical experiments are presented to illustrate our theoretical findings.