Calibration of households preference for open-spaces from an urban cellular automata model: method and application to Dijon

We propose a calibration method for a residential growth model that is grounded on a microeconomic cellular automaton. This model can be seen as a spatial and dynamic representation of an urban economic model with neighbourhood externalities. A 2D spatial equilibrium of residential locations is obtained stepwise through time instead of being a one-shot instantaneous equilibrium. Previous research work showed that the model can produce more or less dense and fragmented urban patterns depending on the preference of households for open space. We propose here a method for calibrating those preferences from simulations of the model and land rents observed within the Dijon urban area. More precisely, equilibrium properties of the model are used to derive the elasticity of open-space preferences. Then simulations are used to estimate the neighbourhood distance to which open-space are valued by households. Our first results tend to support our residential behaviour assumptions and tend to be in accordance with results obtained with more classic methods. Further methodological improvements are however needed.