Research On Workflow Scheduling Simulator And Algorithms Based On Multiprocessor Environment

With the development of multi-core processors and Cloud Computing,people begin to realize the importance of scheduling algorithms.Nowadays,scheduling algorithms applied to single processor are mature,it is widely used in various fields.However,if we use these scheduling algorithms to multi-resource environments directly for convenience,the final result may not be expected.To give researchers a tool to help them analyze problems,we try to build a simulator named VWGSim which integrates workflow generation,data analysis and results display.VWGSim mainly has two core parts:one is the workflow generation under the specified parameters,and the other is the visual display of the scheduling results.Researchers can compare their scheduling results with the optimal scheduling scheme to find the direction of algorithm improvement.With the help of VWGSim,we have done further research on job scheduling algorithm in multi-resource environment.For the two levels of this problem,we propose three effective algorithms.At the job level,we propose MCSW algorithm,it combines jobs which are submitted at the same time to achieve fairness.At the task level,we propose DCBF algorithm and DCMG algorithm.The DCBF algorithm aims to find the location with the maximum task length and resource ratio without violating the task deadline.DCMG algorithm aims to find the position with the smallest gap between the previous task and the the earliest execution time of the task without violating deadline.All the test cases in this paper are generated by VWGSim.These test cases are used to compare MCSW algorithm,DCBF algorithm and DCMG algorithm with their comparison algorithms.It is proved that all of them have good processor utilization and success rate at the corresponding scheduling level.