Research On Task Scheduling Optimization Algorithm In Cloud Environment

Cloud computing is a rapidly developing computing model.It provides various basic services to the world through the Internet,not only facilitates the needs of users,but also promotes the development of emerging company.However,the cloud platform needs to process a large amount of data and computing tasks.How to properly allocate cloud computing resources and design an efficient task scheduling strategy to minimize costs,improve resource utilization,and meet user needs are the key issue to solve in cloud computing.The main contents of this paper are as follows:Firstly,when the genetic algorithm deals with the independent task scheduling in the cloud environment,there are some problems which are slow convergence rate and unstable optimization.Considering the total execution time,execution cost and the imbalance value of the task's load as the optimization goal,a Dynamic Object based on Genetic Algorithm and Particle Swarm Optimization algorithm(DO-GAPSO)is proposed.At the initial stage of population,the Max-Min,the Round-Robin and the Min-Min algorithm are used to provide general direction for the evolution and search of algorithm.In order to improve the convergence speed of the algorithm,the dynamic allocation strategy of linear weight is introduced in the the fitness evaluation modeling.In the selection operation,the individual fitness of the population is calculated,and different selection methods are used to select and ensure the diverse fitness of the population;In the cross operation,the cross-selection mechanism is established to select different crossover mechanisms based on cross–individual situations,While ensuring the stability of the population,the new individuals after the intersection toward a high degree of fitness.In the mutation operation,the particle swarm algorithm is introduced to reconstruct the mutation operator by using the current optimal solution and the historical optimal solution of the population to avoid the blindness of the mutation and guarantee the stability of the population.The optimization ability of the algorithm is enhanced.Secondly,when the traditional algorithm deals with the associated task scheduling in the cloud environment,there are some problems which are Poor performance and the optimization solution cannot meet the diversity needs of users,this paper simulates the process of heuristic algorithm which is the initialization,fitness assessment,task scheduling and selection stages to construct a Hierarchical Evaluation and Dynamic Selection Model(HEDSM).In the initialization phase,in order to ensure that tasks have a certain priority,the workflow task model is preprocessed using the traditional table scheduling algorithm.In the fitness assessment phase,in order to meet the need of two aspects,the difficult evaluation model is constructed from cloud users and cloud service providers.In the task scheduling phase,two-step scheduling is set.First,the policy set is setting,the task is pre-scheduled to ensure that the pre-scheduling scheme inherits the scheduling advantages of each strategy.Second,in order to enhance the performance of the algorithm,the task migration policy is setting to process the pre-scheduling plan.In the selection phase,the appropriate scheduling scheme is dynamic selected according to the different evaluation mode in the set of generated programs.Finally,the paper carries out the compares and analyzes on the Workflow Sim platform.The results show that the DO-GAPSO algorithm improves the convergence speed of the algorithm and ensures the stability of the algorithm optimization based on the completion time,cost and load imbalance of the task set.In addition,the HEDSM model ensures the resource idle rate of the system and improves the task completion time and completion cost,what's more,it is more suitable for task scheduling problems in complex and variable cloud environments.