Two refreshing views of fluctuation theorems through kinematics elements and exponential martingale (Q548107)

General Markov stochastic processes form an integral part of modeling dynamics in statistical mechanics. Usually, they provide a sufficiently realistic description of experimental situations and have traditionally served as a playground for both theoretical and numerical studies. In the present paper the authors were able to revisit, generalize and unify the fluctuation relations (FR) and the generalized fluctuation-dissipation theorem (GFDT) by the language of the kinematics of a general Markov process. The formulation of the fluctuation-dissipation theorem (FDT) in terms of Markov processes is well known in physics but it has a strictly mathematical formulation. For the FR, the fluctuation results were proved for some Markov processes, but the extension to FR for general Markov process is still under debate. The paper consists of seven sections and six appendixes. After the preliminaries in Section 1, we have the necessary notations (e.g. transition probability, Markov generator, stationary and equilibrium state) and definitions (cotransition probability, cogenerator, current and velocity operator) given in Section 2. Section 3 develops the kinematics of a Markov process by defining a set of local derivatives and local velocities associated with such processes. These derivatives appear naturally in the time derivative of a correlation function which appears in the FDT. Section 4 shows the investigations of the form of the kinematics elements, local velocities and velocity operator for some examples of the Markov process which appear in physics: a pure jump process (Section 4.1), a diffusion process (Section 4.2), a mix of diffusion, random jumps and deterministic motion given by a stochastic equation with Gaussian and Poissonian white noises. In Section 5, the reader can find novel results -- the study of the behavior of a Markovian system under perturbation: the notion of the response function to an arbitrary perturbation is given in Section 5.1, a special family of perturbation called Hamiltonian ones or generalized Doob \(h\)-transforms are given in Section 5.2. Section 5.3 proves in a very simple way that the recent GFDT are obtained in this general Markovian context for the case of a Hamiltonian perturbation (there is also a numerical verification). In the next section we have global versions of GFDT which involve a family of functionals called exponential martingales. Section 7 shows the conclusions. The successive appendixes show the proofs of some important equations.