A general procedure for stochastic modelling of systems consisting of a large number of interacting components

Multi‐component systems with interaction are widespread in all the natural sciences and beyond. In this paper, we discuss a general two‐stage procedure for modelling these systems. The procedure is based on first identifying the main influences on incremental system evolution in some phase space and capturing these in a parametric stochastic process structure. The remaining influences are then used for locally fine‐tuning the parameters of these processes. The procedure is discussed in the context of particle sedimentation in viscous fluids and is also applied to the study of interregional migration in population demography. Other applications are indicated. © 1998 John Wiley & Sons, Ltd.