Application of the Maximum (Information) Entropy Principle to Stochastic Processes far from Thermal Equilibrium

Abstract

This paper shows how the maximum (information) entropy principle allows the derivation of the short-time propagator from experimental data provided the process is Markovian. From the propagator, the Fokker-Planck equation can be derived. The Lagrange parameters that are used in the maximum information entropy principle can be derived by minimizing the Kullback information.