Calibration Of Microscopic Traffic Simulation Model For Weaving Sections Of Expressway

At present, the microscopic traffic simulation model has been widely used in the evaluation of traffic operations and management alternatives for urban expressways. However, due to the level of complexity of traffic characteristics, especially the heavy lane changing and weaving behaviors in the weaving sections on urban expressways, it is extremely difficult for a traffic simulation model to capture their characteristics, which greatly affects the accuracy of the simulation results. Further, few of existing studies have focused on the calibration of parameters particularly for the weaving sections of expressways. Therefore, this thesis intends to study the calibration approach of the microscopic traffic simulation model for weaving sections of urban expressways.First, the thesis designs a basic framework of the calibration method for the microscopic traffic simulation model for weaving sections of expressways on the basis of a comprehensive synthesis of state-of-the-art in microscopic traffic simulation calibrations. This framework includes two components:the network model calibration and the traffic model calibration. The later one also includes the calibration of the overall network and the calibration of the local weaving sections, which is the emphasis of the thesis.Then, the thesis chooses the weaving sections on the Western 3rd Ring-road Expressway in Beijing as the study network, analyzes its traffic characteristics based on the field surveys, carries out the sensibility analysis of traffic parameters on the VISSIM simulation platform, and determines the reasonable ranges of the selected calibration parameters, including the desired speed distributions and eight driver behavior parameters.Subsequently, the thesis develops a weaving section calibration model, a multi-objective programming model, considering both the overall network level and the local weaving section level. The speed and traffic count are selected as the indicators of the model performance measurement. Further, the study develops a calculation algorithm based on the Genetic Algorithm (GA), which is programmed using MATLAB and the.NET-based C# languages. Finally, a case study on the road section from Zizhu Bridge to Xinxing Bridge of the Western 3rd Ring-road Expressway is carried out. It uses the proposed calibration method to calibrate the traffic parameters of the simulation model for the network, especially for the weaving sections. It is found that the relative errors between the simulated results and the actual data at all the detectors are less than 15 percents, which validate the feasibility and validity of the proposed calibration method in the thesis.