Fluctuation relations for diffusion processes (Q2517943)

The main goal of the present paper is to propose a unified approach to the derivation of most of the known fluctuation relations in stochastic processes, that is to say the robust identities related to the statistics of entropy production in closed and open systems. To this end, two main ideas are exploited: The first one is the use of time-reversed processes; and the second one is to obtain fluctuation relations by considering diffusion processes generated from a basic original one. The framework involves the Langevin stochastic differential equation, the Stratonovich stochastic differential equation and its forward and backward processes, and the Jarzynski equality together with the Speck-Seifert equality. Some results are stated related to the linearized processes generated by close trajectories.