Konstantin A. Rybakov
Abstract:The paper gives a new representation for the fractional Brownian motion that can be applied to simulate this self-similar random process in continuous time. Such a representation is based on the spectral form of mathematical description and the spectral method. The Legendre polynomials are used as the orthonormal basis. The paper contains all the necessary algorithms and their theoretical foundation, as well as the results of numerical experiments.
What problem does this paper attempt to address?
The main problem that this paper attempts to solve is to provide a new representation method for simulating fractional Brownian motion (fBm), which is a stochastic process showing self - similarity in continuous time. Specifically, the paper aims to represent and simulate fractional Brownian motion through mathematical descriptions in spectral form and spectral methods. The following are the specific contents of the problem:
1. **Deficiencies of Existing Methods**:
- Statistical analysis of data in many practical problems shows that traditional stochastic processes (such as Brownian motion, Poisson process or Lévy process) are not sufficient to fully describe these phenomena.
- Fractional Brownian motion is a generalized Brownian motion and is suitable for describing self - similar stochastic processes, but its simulation and representation methods still need to be improved.
2. **Goals of the New Method**:
- Propose a new representation method based on mathematical descriptions in spectral form to simulate fractional Brownian motion more accurately.
- Use Legendre polynomials as an orthogonal basis to construct this new representation method.
- Realize the expansion of the kernel function \( k_H(t, \tau) \) through matrix algebraic operations and apply it to the simulation of fractional Brownian motion.
3. **Theoretical Basis and Algorithms**:
- The basic definitions and properties of fractional Brownian motion are studied, including its covariance function \( R_H(t, \tau) \) and impulse response function \( k_H(t, \tau) \).
- Using spectral methods, the input and output signals are represented as spectral features, and the system itself is represented as two - dimensional spectral features.
- Stable algorithms for calculating spectral features are developed, and numerical experiments are carried out to verify the theoretical results.
4. **Application Areas**:
- Fractional Brownian motion and the stochastic processes driven by it are widely used in fields such as computer network traffic, weather forecasting, turbulence, geophysics, and financial mathematics.
- The new representation method can improve the modeling and prediction accuracy in these fields.
In summary, the main goal of this paper is to improve the simulation and representation of fractional Brownian motion by introducing a spectral representation method based on Legendre polynomials, so as to be better applied to various practical problems. This method not only provides theoretical innovation but also has important practical application value.
### Formula Summary
- Covariance function of fractional Brownian motion:
\[
R_H(t, \tau) = \frac{t^{2H} + \tau^{2H} - |t - \tau|^{2H}}{2}
\]
- Expression of the kernel function \( k_H(t, \tau) \):
\[
k_H(t, \tau) = a_H (t - \tau)^{H - \frac{1}{2}} {}_2F_1\left(\frac{1}{2} - H, H - \frac{1}{2}, H + \frac{1}{2}, 1 - \frac{t}{\tau}\right) 1(t - \tau)
\]
where,
\[
a_H = \sqrt{\frac{2H \Gamma(H + \frac{1}{2}) \Gamma(\frac{3}{2} - H)}{\Gamma(2 - 2H)}} \quad \text{or} \quad a_H = \sqrt{\frac{\pi H (1 - 2H)}{\Gamma(2 - 2H) \cos(\pi H)}}
\]
- Definition of spectral features:
\[
\Phi_i = \langle q(i, \cdot), \varphi(\cdot) \rangle_{L^2(T)}, \quad i = 0, 1, 2, \ldots
\]
\[
K_{ij} = \langle q(i, \cdot) \otimes q(j, \cdot), k(\cdot) \rangle_{L^2(T^2)}, \quad i, j = 0, 1, 2, \ldots
\]
- Fraction