On characterization of reversible Markov processes by monotonicity of the fluctuation spectral density
Yong Chen.,Minping Qian,JianSheng Xie
DOI: https://doi.org/10.1063/1.2338763
IF: 1.469
2006-01-01
Journal of Mathematical Physics
Abstract:For the system driven by a stationary Markov process {xi(t):t is an element of R+}, if it is in nonequilibrium steady state (i.e., irreversible), then there exists a function phi such that the fluctuation spectrum (or say: power spectrum density) of {phi(xi(t))} is nonmonotonic in [0,+infinity) under mild conditions, which means that there exist nonzero spectrum peaks of the fluctuation spectrum. For the system driven by a Markov chain with discrete time {xi(t):t is an element of Z(+)}, even if it is in equilibrium state (i.e., reversible), one cannot distinguish the equilibrium and nonequilibrium steady state in terms of the monotonicity of the fluctuation spectrum any more.