$\epsilon$-Strong Simulation of Fractional Brownian Motion and Related Stochastic Differential Equations

TL;DR

This paper presents an exact simulation method for fractional Brownian motion and related stochastic differential equations, achieving arbitrary accuracy and enabling efficient, sequential error-controlled simulations.

Contribution

It introduces a novel algorithm for simulating fBM and SDEs driven by fBM with guaranteed error bounds, supporting sequential updates and integration with other methods.

Findings

Achieves almost sure uniform approximation within epsilon for fBM

Extends to SDEs driven by fBM under mild conditions

Supports sequential error updates for efficient simulations

Abstract

Consider the fractional Brownian Motion (fBM) $B^H=\{B^H(t): t \in [0,1] \}$ with Hurst index $H\in (0,1)$. We construct a probability space supporting both $B^H$ and a fully simulatable process $\hat B_{\epsilon}^H $ such that $$\sup_{t\in [0,1]}|B^H(t)-\hat B_{\epsilon}^H(t)| \le \epsilon$$ with probability one for any user specified error parameter $\epsilon>0$. When $H>1/2$, we further enhance our error guarantee to the $\alpha$-H\"older norm for any $\alpha \in (1/2,H)$. This enables us to extend our algorithm to the simulation of fBM driven stochastic differential equations $Y=\{Y(t):t \in[0,1]\}$. Under mild regularity conditions on the drift and diffusion coefficients of $Y$, we construct a probability space supporting both $Y$ and a fully simulatable process $\hat Y_{\epsilon}$ such that $$\sup_{t\in [0,1]}|Y(t)-\hat Y_{\epsilon}^H(t)| \le \epsilon$$ with probability one. Our algorithms enjoy the tolerance-enforcement feature, i.e., the error bounds can be updated sequentially. Thus, the algorithms can be readily combined with other simulation techniques like multilevel Monte Carlo to estimate expectation of functionals of fBMs efficiently.