城市交通流预测关键技术研究

发布时间：2018-03-09 09:17

本文选题：交通流预测　切入点：混沌算法　出处：《西南科技大学》2015年硕士论文　论文类型：学位论文

【摘要】：随着科学与技术的发展,交通拥塞、交通安全和交通环境污染等问题已引起许多科学工作者的广泛关注,交通流系统具有时变、耦合、非线性动态特征。其复杂性、可视化、结构和功能之间的相互关系与内在机理尚未得到足够的重视。故本文对交通流的关键技术进行了研究,其主要内容如下:本文主要分析海量交通流的数据特性以及城市智能交通网中交通流的时空相关性。并基于复杂网络的研究剖析了影响城市交通流变化的主要因素,为交通流预测建模提供决策支持。结合城市交通流特性,利用大数据挖掘技术进行数据的提取、处理等,挖掘其潜在有用的知识为交通流预测建模提供数据支持。一方面,根据当前的交通流数据来动态调整其对未来预测的影响;另一方面,通过对历史交通流数据的时空特性分析,利用数据挖掘领域的相关知识寻求与当前交通流特性最为相似的假设规则。该方法对大数据进行了约简,提取了主要数据,挖掘了未知信息,弥补了传统数据处理在逼近性能方面的不足。同时兼顾交通流预测精度与交通流动态变化的强适应性。研究大规模路网交通流预测,主要通过两种方法建立交通流预测模型(即:基于知识免疫的模型、基于混沌知识免疫的模型)。基于知识免疫的算法是利用知识建构和免疫优化算法的优势相互取长补短,而基于混沌的知识免疫算法是利用混沌算法的优势和知识免疫智能算法内部策略来相互取长补短,从而动态地控制和调节区域大小进而加速其算法搜索速度,以实现高效短时交通流的预测,通过仿真实验证明该方法的可行的和有效的。
[Abstract]:With the development of science and technology, traffic congestion, traffic safety and traffic environmental pollution have attracted the attention of many scientists. Traffic flow systems have the characteristics of time-varying, coupling, nonlinear dynamics, complexity, visualization, etc. The relationship between structure and function and its intrinsic mechanism have not been paid enough attention to, so the key technology of traffic flow is studied in this paper. The main contents are as follows: this paper mainly analyzes the data characteristics of mass traffic flow and the temporal and spatial correlation of traffic flow in urban intelligent transportation network. Based on the research of complex network, it analyzes the main factors that affect the change of urban traffic flow. Based on the characteristics of urban traffic flow, big data mining technology is used to extract and process data, mining its potentially useful knowledge to provide data support for traffic flow forecasting modeling. According to the current traffic flow data to dynamically adjust its impact on the future prediction; on the other hand, by analyzing the space-time characteristics of historical traffic flow data, Using the relevant knowledge in the field of data mining to find the hypothetical rules which are most similar to the current traffic flow characteristics, this method has reduced big data, extracted the main data and mined the unknown information. It makes up for the shortcoming of the traditional data processing in approaching performance. At the same time, it takes into account the strong adaptability of traffic flow prediction precision and traffic flow state change, and studies the traffic flow prediction of large-scale road network. The traffic flow prediction model is established by two methods (i.e., the model based on knowledge immunity, the model based on chaos knowledge immunity). The algorithm based on knowledge immunity is to make use of the advantages of knowledge construction and immune optimization algorithm to complement each other. The knowledge immune algorithm based on chaos makes use of the advantages of chaos algorithm and the internal strategy of knowledge immune intelligent algorithm to complement each other, and dynamically controls and adjusts the size of the region and accelerates the search speed of the algorithm. The simulation results show that the method is feasible and effective.
【学位授予单位】：西南科技大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U491.14

【相似文献】