Abstract:This paper deals with asymptotic errors, limit theorems for errors between numerical and exact solutions of stochastic differential equation (SDE) driven by one-dimensional fractional Brownian motion (fBm). The Euler-Maruyama, higher-order Milstein, and Crank-Nicolson schemes are among the most studied numerical schemes for SDE (fSDE) driven by fBm. Most previous studies of asymptotic errors have derived specific asymptotic errors for these schemes as main theorems or their corollary. Even in the one-dimensional case, the asymptotic error was not determined for the Milstein or the Crank-Nicolson method when the Hurst exponent is less than or equal to $1/3$ with a drift term. We obtained a new evaluation method for convergence and asymptotic errors. This evaluation method improves the conditions under which we can prove convergence of the numerical scheme and obtain the asymptotic error under the same conditions. We have completely determined the asymptotic error of the Milstein method for arbitrary orders. In addition, we have newly determined the asymptotic error of the Crank-Nicolson method for $1/4<H\leq 1/3$.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is: for one - dimensional stochastic differential equations (SDEs) driven by fractional Brownian motion (fBm), study the asymptotic error between the numerical solution and the exact solution and its limit theorem. Specifically, the paper focuses on exploring the convergence and asymptotic error of numerical methods such as Euler - Maruyama, high - order Milstein, and Crank - Nicolson under different Hurst exponents \(H\).
### Main problems:
1. **Determination of asymptotic error**: For stochastic differential equations driven by fractional Brownian motion, how to determine the asymptotic error between the numerical solution and the exact solution?
2. **Convergence conditions**: Under different Hurst exponents \(H\), which numerical methods can ensure convergence, and what are their asymptotic errors?
3. **Exploration of unresolved problems**: In particular, when the Hurst exponent \(H\leq1/3\), the asymptotic errors of the Milstein method and the Crank - Nicolson method have not been fully determined, and the paper attempts to fill this gap.
### Research background:
- Fractional Brownian motion (fBm) is a non - Markov process with long - memory characteristics (when \(H > 1/2\)) or short - memory characteristics (when \(H < 1/2\)). It is widely used to simulate non - Markov time - series data.
- For standard stochastic differential equations driven by standard Brownian motion, there are well - developed theories and numerical methods. However, for SDEs driven by fractional Brownian motion, since fBm is not a semimartingale, traditional numerical methods are no longer applicable.
- The Euler - Maruyama, Milstein, and Crank - Nicolson methods are the most commonly used numerical methods, but their performance under different Hurst exponents is different, especially in the case of \(H\leq1/3\).
### Main contributions of the paper:
- Proposed a new evaluation method to improve the convergence conditions of numerical schemes and obtain asymptotic errors under the same conditions.
- Completely determined the asymptotic error of the Milstein method of any order.
- Newly determined the asymptotic error of the Crank - Nicolson method in the range of \(1/4 < H\leq1/3\).
### Formula representation:
- The covariance function of fractional Brownian motion is:
\[
\text{Cov}(B(s), B(t)) = E[B_sB_t]=\frac{1}{2}(|s|^{2H}+|t|^{2H}-|t - s|^{2H})
\]
- The form of SDE is:
\[
dY_t=\sigma(Y_t)dB_t + b(Y_t)dt
\]
- Euler - Maruyama scheme:
\[
\hat{Y}_t^{(m)}=\hat{Y}_{\tau_r^{(m)}}^{(m)}+\sigma(\hat{Y}_{\tau_r^{(m)}}^{(m)})\delta B_{\tau_r^{(m)},t}+b(\hat{Y}_{\tau_r^{(m)}}^{(m)})(t-\tau_r^{(m)})
\]
- Crank - Nicolson scheme:
\[
\hat{Y}_t^{(m)}=\hat{Y}_{\tau_r^{(m)}}^{(m)}+\frac{1}{2}(\sigma(\hat{Y}_{\tau_r^{(m)}}^{(m)})+\sigma(\hat{Y}_t^{(m)}))\delta B_{\tau_r^{(m)},t}+\frac{1}{2}(b(\hat{Y}_{\tau_r^{(m)}}^{(m)})+b(\hat{Y}_t^{(m)}))(t-\tau_r^{(m)})
\]
Through these studies, the paper provides an important theoretical basis and numerical tools for understanding and applying stochastic differential equations driven by fractional Brownian motion.