Calibration Procedures and Problems of Stability in Nonlinear Dynamic Spatial Interaction Modeling
The use of nonlinear dynamic spatial interaction models is one of the main approaches for the treatment of those phenomena of discontinuity and irreversibility which characterise the complexity of urban evolution. In this paper some methodological and operational problems which arise when a model of this kind, namely the Harris and Wilson model, is calibrated are dealt with. The main aspect of the calibration dealt with here is the problem of the separability of dynamics with different speeds (such as dynamics of stocks and dynamics of activities) when suitable data are lacking. It is shown, among other things, that a fundamental role is played by the intrinsic stability (or instability) of the equilibria of the model. The argument is illustrated by a set of simulations carried out on the Rome metropolitan area, relating to the retailing sector.