Reducing dimensionality and complexity in fine‐grain urban models

A challenge in the ongoing development of mathematical and computer modelling of cities is to add detail, and this implies finer scales—space, sectors and time. This involves specifying high-dimensional arrays and hence large numbers of variables for comprehensive models. The objective of this paper is to show how such large numbers can be reduced by defining sets of principal flows, along with constructing tiers of zone systems, in such a way that computation can be handled at fine scales. The argument is illustrated with the retail model but can easily be extended to a variety of applications.