基于动态数据驱动的交通在线决策

Traffic System On-Line Decision-Making Based on DDDAS

针对交通在线决策问题进行研究,首先介绍了交通仿真模型,分别对宏观、中观、微观交通模型用于在线决策的适用性进行了分析,提出了在线决策所需的交通模型应该具备的特点。基于动态数据驱动思想,对交通在线决策进行了构思。提出了离线计算和在线计算相结合的在线决策系统结构,通过用交通系统历史数据和实时测量数据共同校正仿真系统,实现在线决策系统对实时交通数据的动态整合。并根据动态交通模型特点,提出了与在线校正过程相适应的滚动执行模式。描述了交通在线决策的总体框架,明确了需要研究的关键技术,为交通在线决策系统的实现奠定了基础。

This paper studies the problem of traffic system on-line decision-making. Firstly,we introduced the traffic simulation model,analyzed the applicability of macroscopic,mesoscopic and microscopic models for on-line decision-making separately,and put forward the characteristics of the traffic model for on-line decision-making. And then,the on-line decision-making was conceived based on DDDAS. The system structure of on-line decision-making combining offline computing and online computing was proposed,which realized the dynamic integration of realtime traffic data by using historical traffic data and real-time measurement data to calibrate simulation system. According to the characteristics of dynamic traffic model,the rolling execution mode corresponding to the online calibration process was presented. Our study describes the framework of traffic system on-line decision-making and makes clear the key technologies in it,which lays a foundation for the realization of on-line decision-making system.