ABSTRACT
A research project undertaken after the integrated model package project involved the comparison of then operational urban residence–location models. The essence of the residence–location process as represented in these models centers on the work trip or a combination of the work trip with measures of attractiveness of the potential residence locations. Implicit, therefore, in any estimates of residence locations made by these models, is a set of work–trip estimates. Most of the work done in the UK with residence–location models developed from entropy–maximizing procedures included no attractiveness variables at all. If the calibrations of the equations have yielded reasonably good fits of the model to the base year data, then this should give equally good forecasts. The development of a residence–location model involving a multivariate, multi parameter attractiveness measure plus a two–parameter trip function required the development of new calibration procedures as well.
