Temporal Entropy Evolution in Stochastic and Delayed Dynamics

Abstract

We review the behaviour of the Gibbs' and conditional entropies in deterministic and stochastic systems and continue to a formulation appropriate for a stochastically perturbed system with delayed dynamics. The underlying question driving these investigations: ``Is the origin of the universally observed unidirectionality of time in our universe connected to the behaviour of entropy?'' We focus on temporal entropic behaviour with a review of previous results in deterministic and stochastic systems. Our emphasis is on the temporal behaviour of the Gibbs' and conditional entropies as these entropies give equilibrium results in concordance with experimental findings. In invertible deterministic systems both entropies are temporally constant as has been well known for decades. The addition of stochastic perturbations (a Wiener process) leads to an indeterminate (either increasing or decreasing) behaviour of the Gibbs' entropy, but the conditional entropy monotonically approaches equilibrium with increasing time. The presence of delays in the dynamics, whether stochastically perturbed or not, leads to situations in which the Gibbs' and conditional entropies evolution can be oscillatory and not monotone, and may not approach equilibrium.