Research And Application Of Task Scheduling Algorithm For Police Data Processing

With the advancement of police informationization,the police sysytem has obtained more diverse channels for obtaining data,and the police data obtained has become more and more abundant.It is of great importance to process these police data.It can not only assist police officers to quickly handle crimes,but also plays an important role in various aspects of social life such as maintaining public security and congested traffic.Unluckily the police system is not effective in handling police data.How to effectivly achieve data processing by rationally assigning tasks still faces many problems.For the task scheduling problem of police data processing,this paper mainly made the following research:(1)A cascaded LSTM(Long Short-Term Memory)data prediction strategy is proposed.Because the police data and the data processing machine are separated,the tasks must wait for the arrival of the data before it can be executed.Through the data prediction,the predited data can be processed and sent to the machine in advance and localized on the machine to avoid waiting for data.Take advantage of LSTM,the current data can be predicted based on historical task data.Table and field need to be predicted at the same time,separare training will lose correlation.The cascaded LSTM can predict without losing correlation.The judgment criteria is set for the prediction.Experiments show that this stategy can predict accurately and the efficiency could be improved through the prediction.(2)A task scheduling strategy for police data processing is proposed.In order to improve the execution efficiency and reduce the task failure rate,a task scheduling strategy based on artificial bee colony algorithm and a task reallocation optimization strategy with minimum cost are proposed.The artificial bee colony algorithm is applied to the task scheduling because of its advantages of good searching effect and strong robustness.It transforms the solution of continuous space problem into discrete optimal solution,and learn form the global optimal bee which accelerates convergence.Since the task scheduling strategy based on artificial bee colony algorithm is a static allocation methods,the faliure of allocation may result in failure of the task.The failed tasks can be reallocated by moving the assigned tasks.In order to reduce the impact of redistribution on follow-up tasks,a cost formula is proposed to select the task with the least cost to move and reduce the task failure rate.It is proved that the task scheduling strategy can improve effeciency and reduce task failure rate through experiments.(3)Designed and implemented a multi-layer high-performance distributed security data service platform.The overall structure and business process of the platform are designed,and the architecture and interaction of the computing center module,requisition center module and data center module are described in detail.The data prediction strategy based on the cascaded LSTM and the task scheduling strategy for police data processing are used into relevant module.Finally the police data processing results are displayed in the platform.