Hybrid scheme for Brownian semistationary processes (Q6032782): Difference between revisions

The Brownian semistationary \((\mathcal{BSS})\) processes form a flexible class of stochastic processes that are able to capture some common feature of empirical time series, such as stochastic volatility, roughness, stationarity and strong dependence. In the present paper a simulation scheme for Brownian semistationary processes, which is based on discretizing the stochastic integral representation of the process in the time domain, is introduced. It is assumed that the kernel function of the process is regularly varying at zero. The kernel function is approximated by a power function near zero and by a step function elsewhere. The resulting approximation of the process is a combination of Wiener integrals of the power function and a Riemann sum, which is called a hybrid scheme. The asymptotic of the mean square error of the hybrid scheme is described. The hybrid scheme is tested by two numerical experiments. In the introduction of the paper the concept of a \(\mathcal{BSS}\) process is described. A \(\mathcal{BSS}\) process \(X\) is defined by the integral representation \[ X_t = \int_{-\infty}^t g(t-s)\sigma_s d W_s, \] where \(W\) is a two-sided Brownian motion, \(\sigma\) is a stochastic volatility and \(g\) is the kernel function. To the end of the introduction some known methods for the simulation of process \(X\) are discussed. In Section 2 the rigorous definition of a \(\mathcal{BSS}\) process is recalled. Some assumptions concerning the kernel function \(g\) are introduced. Also, the concept of the hybrid scheme is introduced. The so-called gamma-kernel and the power-law kernel are considered. Theorem 2.5 is the main theoretical result. Under introduced assumption, a sharp description of the asymptotic behavior of the mean square error of the hybrid scheme is shown. The hybrid scheme is extended to a class of non-stationary processes. The truncated Brownian semistationary \((\mathcal{TBSS})\) process of the form \[ Y_t = \int_0^tg(t-s) \sigma_s d W_s, \;\;t \geq 0, \] is considered. In Theorem 2.12 the asymptotic behavior of the mean square error of the hybrid scheme for the \(\mathcal{TBSS}\) process is obtained. In Section 3 the implementation of the discretization scheme is discussed. Some numerical experiments are presented. The change-of-frequency statistics and the rough Bergomi model of Bayer are considered. Section 4 contains the proofs of the theoretical and technical results given in the paper.