{"id":"arxiv/2601.21428","url":"https://arxiv.org/abs/2601.21428","tags":["math.NA"],"score":0,"title":"Numerical Methods for Dynamical Low-Rank Approximations of Stochastic Differential Equations -- Part I: Time discretization","domain":"math.NA","source":"arxiv","authors":["Yoshihito Kazashi","Fabio Nobile","Fabio Zoccolan"],"reviews":0,"subject":"math.NA","summary":"In this work (Part I), we study three time-discretization schemes for the Dynamical Low-Rank Approximation (DLRA) of high-dimensional stochastic differential equations (SDEs). Specifically, we consider the Dynamically Orthogonal (DO) method for DLRA proposed and analyzed in arXiv:2308.11581v4, which approximates the true solution by a linear combination of few products between deterministic orthonormal modes and stochastic modes, both time-dependent. The first scheme considered consists in a forward discretization in time of both deterministic and stochastic components, in a Euler-Maruyama…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-01-29","reactions":{}}