چكيده به لاتين
Micro-Simulation software is a software that models driving behaviors with the help of defined mathematical models. By means of these simulators, the traffic flow and the interaction between vehicles can be analyzed perfectly. In Iran, drivers have non-lane-based behavior; therefore, the dissimilar behavior of drivers between developing and developed countries denies the possibility of using models and results of the current micro-simulation software, which is calibrated based on the lane-based behavior, for traffic studies in Iran.
In this study, in order to have a further understanding of such traffic behavior particularly in traffic congestion, a lane drop, as a bottleneck, was modeled and calibrated in VISSIM using measured data, Design of Experiment and Genetic Algorithm. Results showed that some parameters such as Minimum Lateral Distance were not determined by the Design of Experiment due to the fact that the aforementioned software did not have any sense on driving behavior in the animation of the software. The RMSNE, Root Mean Square Normalized Error which included average travel time and density, was 0.4892 and input flows had less than 15% error. I believe that Ax and Min Look Ahead Distance, (for car-following model) and Keep Lateral Distance to Vehices on the Next Lane(s) were the most important parameters of the VISSIM which had many effects on the results and non-lane-based behavior. All in all, the results can be used to study the traffic in any other bottlenecks for a network with no lane discipline.
