基于Hadoop的交通信息物理系统交通诱导方法研究

AbstractAiming at the characteristics of high throughput data traffic and timeliness,inorder to more efficiently handle this type of data,explore a method to realize service modelbased on cloud computing,traffic data can be extended by Hadoop technology architectureprocessing,services,and implement a prototype system.The main idea of this method isthe parallel processing for distributed file using Hadoop's ability to deal with the massivetraffic data,and to integrate the Flume+Kafka+Spark Streaming data system constructionand processing as the basic points of innovation platform,finally reference A*algorithm,optimization of road network equilibrium assignment,simplified path search method.Thisprocess has higher efficiency and higher reliability,and has more prominent advantagescompared with traditional methods.Around this method the design and implementation ofthe traffic information service system based on Hadoop,according to the actual traveldemand,taking into account the optimal allocation of city traffic flow,traffic guidanceprovide distributed service to users;by automatically triggering Hadoop access accordingto user request calculation task to build a middle layer function,solves the original Hadoopthe off-line batch processing mode is not suitable for real-time online information serviceproblems.Finally,a simulation example is given to demonstrate the effectiveness of thesystem.Test results show that this method can greatly improve the data processingefficiency,achieved satisfactory results,and further demonstrates the reliability of thismethod for the data processing and effectiveness,to ensure the traffic system data flowtimeliness and high tolerance type processing.Keywords:traffic guidance system,Hadoop frame structure,big data systemconstruction,algorithm optimization,processing path planning