| This dissertation focuses on multiclass equilibrium demand synthesis (MEDS) models which attempt to estimate entire freight delivery tours in conjunction with passenger car trips. The procedure is based on entropy maximization considering traffic counts and congestion, reaching an equilibrium in the network.;The amount of research conducted on the estimation of freight tours is extremely small. However, in spite of their value, the assumptions used in these models cannot fully capture the immense complexity of real life delivery tours. As a result, the practitioners and research community do not have access to the tools needed to estimate freight/passenger demand models in situations where financial constraints prevent full scale data collection efforts. The main goal of this research is to help address this need. In doing so, the dissertation pursues two interrelated objectives. The first objective is to develop MEDS travel demand models based on entropy maximization to estimate freight demand using secondary sources (traffic counts) under the presence of other vehicles in order to reach an equilibrium in the transportation network. The second objective is to test the effectiveness of the models and alternative sampling procedures to select/collect the traffic/input for two different case studies (Sioux Falls, SD and Denver, CO) where true demands are known.;The test results show that the estimated tour flows closely matched the observations in the proposed formulation, and the results are better as more information is available. Furthermore, the results clearly indicate that, in order to minimize the assessment error associated with the estimation of freight tours, the traffic counts should be collected such that the largest links with "independent" traffic are surveyed and included as inputs to the models. Finally, the results illustrate that the presence of different vehicle types affects the travel times in the network, where an equilibrium is reached based on the freight tour and passenger car flows. |
