ABSTRACT
In this chapter we discuss a dynamic spatial microsimulation model of a city region. The overall approach is to characterize the population as a series of individuals with specific demographic attributes, in which the core attributes are age, gender, occupation, marital status, health and ethnicity. The individuals are combined into households, and acquire further properties (tenure, housing type) in addition to membership qualities (household size, age and occupation etc. of the head). Populations are characterized using synthetic estimation procedures, so that for any small area the individuals are realistic, but not in any meaningful sense real or identifiable. More detail on the estimation of initial or baseline populations is provided by Harland et al. (2012), while the dynamics are discussed by Wu et al. (2010). The population of quite large cities, or even countries, can be represented quite comfortably in this way using the computational resources now available. For a discussion of potential world models with billions of individual ‘agents’ then see Epstein (2009). Some of the computational issues and infrastructures to support complex processing of such large datasets (specifically, spatially dynamic models) are reviewed by Birkin et al. (2010).
