Extended Thermodynamic Relation and Fluctuation Theorem in Stochastic Dynamics with Time Reversed Process

Abstract

We consider a stochastic model described by two stochastic differential equations of motion; one is for the stochastic evolution forward in time and the other for backward in time. We further introduce averaged quantities for the two processes and construct the extended thermodynamic relation following the strategy of Sekimoto. By using this relation, we derive the fluctuation theorems such as the Seifert relation, the Jarzynski relation and the Komatsu-Nakagawa non-equilibrium steady state with respect to the introduced averaged quantities.