Parameter Calibration of VISSIM Based on Genetic Algorithm

Traffic simulation models are widely used in traffic system management and optimization. Parameter calibration is an important step before using traffic simulation software. In this study, the simulation software VISSIM was taken as an example to study the calibration problem of traffic simulation model, and the method flow of parameter calibration was designed and realized by using genetic algorithm. In view of the problems of genetic algorithm, such as easy to fall into local optimal and slow convergence speed, the initial population and variation of the algorithm were improved, and the algorithm was realized by using MATLAB. By combining experimental data to calibrate it, the convergence iteration times of genetic algorithm before and after improvement are reduced by 37.5%, and the speed improvement effect is significant. Meanwhile, the simulation error of parameter index after calibration based on the improved genetic algorithm is controlled at about 20%, which verifies the effectiveness of the designed method.