A Task Scheduling System Model Based On Parallel Resource Attribute Selection

With the advent of the era of cloud computing and big data, it is a challengingproblem to implement large-scale computing tasks effectively and allocate computingresources to these tasks reasonably. The problem of task scheduling belongs to not onlythe field of computer science but also the field of operations research, artificialintelligence and applied mathematics. Therefore, it has great theoretical significance toresearch and develop an efficient task scheduling system.In order to develop an effective and rational task scheduler, we have conducted alot of research on current scheduling algorithms and systems. Then we have designed ascheduling system and a resource attribute selection algorithm on the basis of theoriginal scheduling framework. The major work of this algorithm includes defining theresource attribute of executors, sending a test task to determine the suitable type of taskfor the executors, designing a fitness function to reduce the matching resource time foreach task based on gene expression programming algorithm. The proposed schedulingmodel improves the performance of the original scheduling framework by means ofimproving the scheduling strategy, building resource pools of executors and addingresource attribute selection algorithm.Finally, the scheduling system based on parallel resource attribute selectionalgorithm has a distinct advantage on the executing time of a certain number of tasks,the throughput and the load level of the fixed executing node in comparison to otherthree scheduling algorithm and scheduling systems. In terms of practical applications,two types of tasks of real-time and off-line processing are tested. From theimplementation point of view, the system model researched in this paper has goodperformance in task scheduling.