The Calibration of Trip Distribution Models

In this paper, model calibration is considered in terms of the evidence supplied by a survey to substantiate the distributional hypothesis implied by the model. By means of Bayes theorem, an evidence test is derived which is applicable to trip-distribution models. The test produces a set of constraint equations to determine optimal values for the model parameters. Section 6 determines the constraint equations necessary to calibrate some commonly used models and indicates how, for comparative purposes, they can be written in a standard format. The final section gives a possible method of computation applicable to the calibration of a variety of models.