Calibration Of VISSIM Microscopic Traffic Simulation Model

:The calibration of traffic simulation model parameters directly affects the output quality of the results, and it is an important guarantee of the validity of the simulation results. It is clear that the selections of model parameters to be calibrated and calibration methodologies are much more important in the overall traffic simulation modeling process. Formly, many of existing studies have focused on the calibration of parameters, particularly for the highways and their weaving sections, and few pay close attention to the calibration of smaller regions, such as the urban intersections. Specifically, there are lack of researches on the model parameters as the same as a simple graphical user interface for a good algorithm, which would make automated calibration much more accessible to modeling engineers. Therefore, this thesis intends to study the calibration approach of the microscopic traffic simulation model for the urban intersections.First, the thesis provides an overview of the microscopic traffic simulation, focusing on the adaptation of the microscopic traffic simulation models and calibration levels. Then a comprehensive synthesis of the microscopic traffic simulation calibrations is given for three levels of the calibration process, calibration parameter sets and the optimization algorithms. According to the model adaptability and calibration levels’requirements, VISSIM is selected.Then, a signal controlled intersection is chosen to analyse the VISSIM microscopic simulation calibration of the evaluation parameters and calibration parameters. The output indexes for the calibration of VISSIM results are summed to determine the queue length and travel time as the evaluation indicators. To give the VISSIM key calibration parameters, the thesis analyses microscopic traffic simulation impact parameters and the VISSIM’s formulas. By the analysis of variance results,4core driving behavior parameters are given.Again, based on the selected simulation evaluation and calibration parameters, the thesis develops a genetic algorithm-based graphical user interface in the calibration procedure. A weighted coefficient is introduced to weigh the differences in treatments of the queue length and travel time accuracy in the fitness function. Then, the genetic operators have been embedded in the VISSIM microscopic traffic simulation parameter calibration system. By using the.NET based VB languages and VISSIM COM component library, the thesis develops a set of objects for the easy automatic calibration procedure for urban intersections.Finally, a case study of the Rong Jing Street intersection in Beijing is carried out. It shows that the instance of the intersection in the case has fast convergence characteristics, and calibration results of the error control about10%calibration accuracy to meet the model requirements.