Kinetical theory beyond conventional approximations and 1/f-noise

Abstract

The 1/f-noise is considered which has no relation to long-lasting processes but originates from the same dynamical mechanisms as what are responsible for the loss of memory, fast relaxation and usual shot noise. The universal long-range statistics of memoryless random flows of events and of related Brownian motion is analysed whose peculiarity is close connection between spectral properties and non-Gaussian probabilistic properties both determined by only short-range characteristic time scales. The exact relations between equilibrium and non-equilibrium 1/f-noise in thermodynamical systems are presented. The existence of long-living higher-order correlations and flicker noise in the Kac's ring model is demonstrated. It is shown that conventional Boltzmannian gas kinetics losses 1/f-noise because incorrectly takes into account the conservation of particles and probabilities in collision events. The correctly performed translation of BBGKY hierarchy into kinetical language, under Boltzmann-Grad limit, results in infinite chain of kinetical equations which describe probability distributions on the hypersurfaces corresponding to encounters and collisions of particles. These equations can be reduced to Boltzmann's equation only in the limit of spatially uniform gas, but generally forbid the molecular chaos. If being applied to Brownian motion of gas particles, the theory yields 1/f fluctuations of self-diffusivity and mobility. This is example of Hamiltonian dynamics which produces 1/f-noise without any slow processes and long life-times.