Abstract
The trip-distribution problem is discrete in formulation and combinatorial in nature. This truism has been ignored, as well as obscured, by the total attention focused on the use of the asymptotic methods of the gravity model for the maximum-entropy estimation of interzonal trips. These estimates appear to be reasonably reliable in indicating the regions—not the exact states—of highest probability. The Darwin–Fowler derivation of the gravity model demonstrates that the asymptotic solution is exactly equal to the average of all feasible microstates. Calibration of the model can be based on this proposition. An approach to testing the principle that all microstates are equally probable is outlined.

This publication has 12 references indexed in Scilit: