Optimization Algorithm For Task Scheduling In Cloud Computing

Task scheduling in cloud computing environment is critical for cloud computing applications and is a key technology for improving the performance of cloud computing systems.Reasonable and effective task scheduling directly affects the load and performance of the system.It is of great significance for improving the utilization of computing resources,reducing operating costs,and meeting user requirements.Firstly,according to the common scheduling strategy of independent tasks,considering the completion time and the completion cost of the task set,as the optimization goal of the algorithm,under the premise of guaranteeing system load balancing,an improved Simulated Annealing Multi-population Genetic Algorithm(SAMPGA)is proposed.The algorithm divides multiple populations into families with different evolutionary directions.Each family adopts different evolutionary strategies to overcome the disadvantages of the traditional genetic algorithm with premature convergence and falling into local optimum.At the initial stage of population,the max-min and min-min algorithm are used to avoid the blindness of the algorithm search.In the genetic operation,the Simulated Annealing(SA)algorithm is introduced,and the next generation is filtered using the Metropolis criterion to ensure the diversity of the population and enhance the global search ability of the algorithm.Secondly,for the case of priority constraints between scheduling tasks,based on the classic list scheduling algorithm and task duplication strategy,an improved list scheduling algorithm called Heterogeneous Earliest Finish Time with Duplication(HEFTD)is proposed.It improves the task priority calculation method in the algorithm.In the process of calculating the priority,considering the influence of the task on the successor task and increasing the foresight,it can construct a reasonable priority task queue;and task duplication is introduced in the idle time slot of the virtual machine.Duplication of tasks reduces waiting time and avoids unnecessary communication costs,thereby increasing the utilization of virtual machines and shortening the scheduling length of related tasks.Finally,this dissertation carries out experiments on CloudSim simulation platform.The experimental results show that the Simulated Annealing Multi-population Genetic Algorithm can effectively reduce the completion time and cost of the task set,and ensure the load balance of the system.In addition,Heterogeneous Earliest Finish Time with Duplication algorithm proposed in this dissertation can effectively reduce the scheduling length of the task set compared with Heterogeneous Earliest Finish Time algorithm,and is an effective and reasonable task scheduling method in the cloud environment.