Research On Scheduling Algorithm Optimization Of Data Cloud Platform For High-speed Railway Catenary Monitoring

China’s high-speed railway operating mileage has nearly 40,000 kilometers,highspeed rail catenary system is the power source of high-speed train operation.At present,the maintenance and operation of catenary mainly adopts manual or catenary inspection car to carry out regular inspection.The research on the related technology of online real-time monitoring system of catenary status is of great significance to improve the intelligent operation and maintenance level of catenary and realize the early warning of catenary faults.Based on the monitoring status and actual functional requirements of the monitoring system for catenary,this paper studies the design and application of data cloud platform for High-speed Railway catenary monitoring,and makes an in-depth study of the Hadoop platform and its scheduling algorithm,and optimizes the scheduling algorithm based on computing power scheduling algorithm.High-speed Railway catenary monitoring data cloud platform server adopts B/S mode design and realizes functions such as real-time data display,historical data query,alarm processing,data analysis and system management.To deal with the massive real-time data of the monitoring system,the database design is carried out.Relational database Mysql,distributed NoSQL database Redis and OpenTSDB are used to build the database system to achieve the efficient storage and management of monitoring data.OpenTSDB is a temporal database based on Hbase,running on top of the Hadoop ecosystem.In order to improve the performance of Hadoop platform resource scheduler in heterogeneous environments,the improved Capacity scheduling algorithm is proposed based on job load types,real-time node loads,and job priority weights,which optimizes the load balancing of the system to a certain extent and improves the resource utilization rate of the cluster.In this paper,data cloud platform for High-speed Railway catenary monitoring and Hadoop test cluster are constructed to verify the functions and performance of the monitoring system and the improved Capacity scheduling algorithm respectively.The test results show that each functional module of the system can run stably and meet the functional requirements of the catenary.Compared with Capacity Scheduler,the improved Capacity scheduling algorithm improves the average completion time of jobs.In the single job type test,the average job completion time is reduced by 10.4% for I/O load type job and 12.0% for CPU load type job.In the mixed job type test,the average job completion time is reduced by 30.8%.