Abstract
Most computable urban models, designed to replicate the locational distribution of socio-economic activities, are based on aggregate patterns of employment, population and spatial interaction. Such structures often called LUTI models can now be built very rapidly at scale so that they can be used to simulate the impact of large-scale economic and demographic change on the way a city or system of cities is able to embrace locational economies of agglomeration. We first explore the way such models can be articulated as simultaneous interactions of employment and population, and we demonstrate how they can be solved iteratively to mirror a system in equilibrium. We apply the model to the CAMKOX Corridor (the Cambridge-Milton-Keynes-Oxford Arc) and empirically investigate the model’s properties, indicating how the model can be used to predict the agglomeration economies of changes in that region, and then illustrating how an array of possible solutions can be generated as different varieties of digital twin. As the aggregate LUTI models we develop can now be run hundreds, if not thousands, of times, we illustrate how sensitivity testing, scenario generation and changes in locational behaviour can now be tested routinely, thus developing simulations that bound define the wider solution space of different model types.
Authors: Michael Batty, Willow Liu and Jens Kandt
Download CASA Working Paper 241 (file size 5 MB, file format pdf)