基于自适应正交遗传算法的道路交叉口VISSIM模型参数标定方法

发布时间：2019-02-09 18:08

【摘要】：交通仿真作为解决交通问题的有力方法,在交通工程领域得到了广泛的应用。然而大多数软件是根据开发国的交通条件研发的,并不适用于我国,需要对模型参数进行标定才能准确表达仿真对象。针对这一现象,本文研究了自适应正交遗传算法在道路交叉口VISSIM模型参数标定中的应用。本文首先简单介绍了微观交通仿真模型参数标定的研究现状,对国内外微观交通仿真模型参数标定优化方法的选取进行了回顾和总结,并指出微观仿真模型参数标定在国内的现实意义。其次,论文以道路交叉口的VISSIM模型为背景,从评价指标的选取、标定参数的选择、敏感参数的识别三个层面,对仿真模型参数进行了介绍,并运用敏感性分析确定了交叉口模型的标定参数集,为后续标定工作奠定了基础。再次,针对传统遗传算法的不足,论文引入自适应正交遗传算法。该方法最大的特点是运用正交试验设计对生成初始种群的方法加以改进,使种群均匀分布于解空间,并引用自适应算法,使交叉算子具有自适应性。然后,基于自适应正交遗传算法设计了道路交叉口VISSIM模型参数标定方法,详细给出了该优化方法中染色体的编码解码、适应度函数和自适应正交交叉算子的设计。借助COM接口和VB程序语言,完成了基于自适应正交遗传算法的道路交叉口VISSIM模型参数标定程序。最后,将基于自适应正交遗传算法的参数标定程序应用到北京市荣华中路与荣京西街交叉口模型中,实现参数的自动化标定,获取更能反映交叉口现状的仿真参数和结果。通过实验的对比分析,说明自适应正交遗传算法的优越性。
[Abstract]:As a powerful method to solve traffic problems, traffic simulation has been widely used in traffic engineering field. However, most of the software is developed according to the traffic conditions of the developing country, and it is not suitable for our country. It is necessary to calibrate the model parameters in order to express the simulation object accurately. In view of this phenomenon, the application of adaptive orthogonal genetic algorithm in the calibration of VISSIM model parameters at road intersection is studied in this paper. In this paper, the research status of parameter calibration of microscopic traffic simulation model is introduced briefly, and the selection and optimization methods of parameters calibration of microscopic traffic simulation model at home and abroad are reviewed and summarized. It also points out the practical significance of parameter calibration of micro simulation model in China. Secondly, taking the VISSIM model of road intersection as the background, the paper introduces the simulation model parameters from three aspects: the selection of evaluation index, the selection of calibration parameters and the identification of sensitive parameters. The calibration parameter set of the intersection model is determined by sensitivity analysis, which lays a foundation for the subsequent calibration work. Thirdly, aiming at the deficiency of traditional genetic algorithm, adaptive orthogonal genetic algorithm is introduced in this paper. The most important feature of this method is that the orthogonal design is used to improve the method of generating the initial population, so that the population is uniformly distributed in the solution space, and the adaptive algorithm is used to make the crossover operator self-adaptive. Then, based on the adaptive orthogonal genetic algorithm, the parameter calibration method of VISSIM model for road intersection is designed. The coding and decoding of chromosome, the fitness function and the design of adaptive orthogonal crossover operator in the optimization method are given in detail. With the help of COM interface and VB program language, the parameter calibration program of road intersection VISSIM model based on adaptive orthogonal genetic algorithm is completed. Finally, the parameter calibration program based on adaptive orthogonal genetic algorithm is applied to the intersection model of Rong Hua Middle Road and Rongjingxi Street in Beijing to realize the automatic calibration of the parameters and obtain the simulation parameters and results which can reflect the status quo of the intersection more effectively. The advantages of the adaptive orthogonal genetic algorithm (AGA) are proved by the comparative analysis of experiments.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U491.23;TP18

【参考文献】