Dynamical representation of Markov processes of the separable class

The mean-squared continuous Markov process of the separable class is represented by a nonlinear stochastic differential equation. The representation for the strictly stationary case implies that the process is determined by its autocorrelation function and first-order probability density function. A class of stationary Markov separable processes may be obtained by a zero-memory nonlinear (ZNL) transformation of a wider class of stationary Markov processes. A special case of the multidimensional process is shown to result in a separable process of degree . Several examples are considered to illustrate the representation.