“…A numerical treatment of the phase space density requires, by necessity, a transformation of the Liouville density function σ (x, t) into a probability, through the relation p (x i , t) = σ (x i , t) δx i (1) where the x-axis is divided into N equal parts indexed by j = 1, 2, ..., N, δx i is a small phase space interval (x i , x i + δx i ) and the probability density is normalized to unity at all times…”