A Markovian Entropy-Maximizing Model of Population Distribution

Abstract

The entropy-maximizing formalism used in urban and regional modelling has typically been applied within a static or equilibrium context. This paper presents a dynamic entropy model of the distribution of population over time. It is initially assumed that a Markov chain adequately represents the residential relocation process. The strategy then involves maximizing the entropy of a Markov chain, subject to suitable constraints, so as to generate least-biased estimates of the Markovian parameters. If a stationary process is assumed, these in turn allow the projection of the probability distribution vector of population densities over successive, equal, time intervals.