Nonstationary

In the non-stationary process of doubling the frequency of high-power laser radiation in metamaterials, the effects of self-and cross-phase modulation, related to the cubic nonlinearity of the medium, are considered. It is assumed that the pump wave propagates in a region with negative refraction. It is shown that in the case of a nonlinear optical process of frequency doubling of an intense light wave in a metamaterial, it is important to take into account the influence of a powerful pump wave. It is shown that the frequency shift of the maximum of the harmonic spectrum in the metamaterial is related to the cubic nonlinearity of the medium. It is shown how the width of the spectrum of the second-harmonic depends on the pump intensity, linear and nonlinear phase detuning. In addition, analytical expressions are obtained for the optimal values of frequency detuning and group velocities on which the maximum of the harmonic spectrum is obtained for non-stationary generation in a metamaterial. Estimations of the quasistatic interaction lengths and the frequency shift of the second-harmonic spectrum are given. By varying the linear phase detuning or the pump intensity, it is possible to manipulate the frequency shift of the maximum of the harmonic spectrum, and from here one can realize a smooth rearrangement of the central frequency of the harmonic spectrum.

This paper proposes a change point detection for electroencephalograms (EEG) signal application based on Particle Swarm Optimization (PSO). As EEG signal is well known consider as non-stationary in nature, we model the signal by using the sinusoidal-Heaviside function, which are capable to represent the change of the behavior of the signal. The parameter of the model with the change point location can be tuned by finding the minimum value of sum squared error. It was showed that the minimum value of sum squared error in the parameter tuning give the exact location of change point. The proposed method is applied to the human EEG during an eye moving task. © 2011 Springer-Verlag.

http://link.springer.com/chapter/10.1007/978-3-642-21729-6_122

In this paper we consider some non stationary relaxed synchronous and asynchronous multisplitting methods for solving the linear complementarity problems with their coefficient matrices being H−matrices. The convergence theorems of the methods are given,and the efficiency is shown by numerical tests.