Derivation of nested transport models within a mathematical programming framework

This paper presents an optimisation framework for deriving transport demand models with a nested choice structure. The mathematical program is based on the separability of the entropy term in the objective function. The resulting models have similar functional forms to those formed from random utility maximisation principles. However, some precautions must be taken as to the conditional sequence chosen, the uneven sizes of choice subsets, and unlikely values of parameters. The required parameters cannot be directly evaluated by simply satisfying a trip cost constraint because the nested model structure implies endogeneous measures to attribute to each level of demand. A computational procedure based on the satisfaction of entropy constraints is presented for calibrating each level-related parameter.