ABSTRACT
J tion which allows these stock variables to be determined within that mathematical programme - so that the pattern which results is really determined only by the given distribution of basic employment, the c .. -matrix and any intrinsic attractiveness
lJ factors. One interesting result to emerge is that the form of the constraints for the linked models, when combined into a single programming model, creates a residential location pattern which is dependent both on access to services as well as the more usual access to jobs, and so this makes the interacting field concept introduced in Section 5.2.4 appear more complicated. It is also possible to continue this sequence of argument in the direction of random utility models: we take one such model, that developed by Coelho and Williams (more precisely a consumers' surplus model which results from such an analysis) as our example.
