An entropy-based modal split model

A mathematical programming model for modal split predictions based on entropy maximization is presented. The relationship between this type of entropy-based model and the more traditional logit-type models is clarified. The entropy model presented has been used to predict travel patterns in the Stockholm region and has been based on census and survey data. A balancing procedure has been used in the solution process. It is shown how additional information, in this case traffic count information, can be included in the model. Computational results and experiences are discussed.