|M.Sc Student||Saji' Elias|
|Subject||Investigation of the Impact Range of Traffic Assignment and|
Simulation Models in Transportation Planning
|Department||Department of Civil and Environmental Engineering||Supervisor||Full Professor Bekhor Shlomo|
|Full Thesis text - in Hebrew|
The traffic assignment model is commonly used to predict traffic. Recent transportation models are based on micro-simulation traffic representation, which take into account individual car movements, as opposed to aggregate traffic flows in the assignment process; thus, traffic flows are obtained by aggregating the individual cars. Travel times obtained in this process are consistent with traffic flows, since the model takes into account queue formation. Accordingly, micro-simulation models require additional data compared to assignment models, and therefore they are generally applied at the local level.
The objectives of this study are: investigate the different characteristics of traffic assignment model and micro-simulation model, classify the pros and cons of every method, testing the impact of every method, and diagnose the models in a real situation.
The proposed method is based on an iterative process of both traffic assignment model and micro simulation model. First, a traffic assignment model for a given volume-delay function is performed at the system level. This model allows the definition of the relevant project region, and creates a demand matrix for this region. This matrix serves as input to the micro-simulation model. Then, the micro-simulation model is applied, giving more realistic travel times in the segments of the project region network. The travel times obtained by the simulation model serve as input to update the volume-delay function in the network segments of the traffic assignment model. This process continues until the trip matrix in the previous iteration is sufficiently close to the present iteration.
The proposed method is examined empirically in a test case - the Hamovil junction located in the Haifa Metropolitan Area. The test case results indicate that there are significant differences between the parameters of the volume-delay function obtained in the system level and in the local level. These differences were found by comparing the parameters from the original assignment model and the calibrated parameters generated by the combined assignment - simulation calibration process. This thesis shows that it is possible to combine the two models.
A possible research direction is to increase the simulation area (more than one junction), and examine the reciprocal effects between junctions and adjacent network links. Another possible extension is the conversion of the static assignment model to a dynamic assignment model and testing the combination of this model and the micro simulation model.