论文部分内容阅读
为使交通仿真模型能够更好地运用于各种实际交通运行方案的测试中,需要对交通仿真模型进行参数标定。在微观交通仿真模型的基础上,提出微观交通参数及动态路径选择的校正设计流程;通过对仿真模型的敏感性分析,选取对参数校正影响较大的5个参数进行研究,然后利用遗传算法进行参数校正,并采用试错法将路径校正纳入整个参数校正流程当中。构建以昆明市一二一大街、学府路为主的区域路网并在Q-paramics软件中仿真模拟,最后选取交叉概率为0.5、变异概率为0.1的仿真因子,搜索出最优参数组合。仿真结果显示,相对误差基本控制在10%左右证明了校正结果的合理性及具体实现方法的有效性。
In order to make the traffic simulation model be better used in the testing of various practical traffic scenarios, the traffic simulation model needs to be parameter calibrated. Based on the micro-traffic simulation model, the calibration design flow of micro-traffic parameters and dynamic path selection is proposed. By analyzing the sensitivity of the simulation model, five parameters that have a great influence on the parameter calibration are selected, and then the genetic algorithm Parameter correction, and the trial and error method will be included in the path correction of the entire parameter correction process. Constructing a regional road network which is mainly composed of 12th Avenue, Xuefu Road in Kunming City and simulating it in Q-paramics software. Finally, we select the simulation factor with a crossover probability of 0.5 and mutation probability of 0.1, and find the optimal combination of parameters. Simulation results show that the relative error of about 10% basic control to prove the rationality of the calibration results and the effectiveness of the specific implementation.