Strong 1.5 order scheme for fractional Langevin equation based on spectral approximation of white noise

Yibo Wang,Wanrong Cao
DOI: https://doi.org/10.1007/s11075-023-01576-z
IF: 2.37
2023-06-29
Numerical Algorithms
Abstract:A full-discrete scheme is presented for a Langevin equation involving Caputo fractional derivative and additive white noise. Based on a spectral truncation of white noise, the fractional Langevin equation is converted to an approximate equation with random parameters and a finite difference scheme is constructed. Consistency of the approximate equation as well as error estimate of the finite difference scheme are obtained. It is proved that when spectral truncation level and the step size are inversely proportional, the convergence order of the finite difference scheme is 1.5 in mean-square sense and independent of the value of fractional derivative order. Numerical examples verify the theoretical analysis. Moreover, to fulfill long-time simulation, a scheme based on piecewise spectral approximation of white noise is further developed.
mathematics, applied
What problem does this paper attempt to address?