Improved CRO Multi-objective Optimization Algorithm For DAG Task Scheduling

In the past few years,cloud computing has become one of the most popular areas of the emerging technology industry,in the cloud computing "as needed,according to the amount of payment" business service mode,parallel task scheduling has become a key technology in the field of cloud computing.Task scheduling problem on cloud is a NP-hard combinatorial optimization problem.The random search technique based on swarm intelligence has the potential to solve task scheduling problem,and a better solution is obtained.However,most of the existing cloud task scheduling algorithms is relatively simple.Therefore to establish the new problem model based on DAG(Directed A-cyclic Graph)model and design a new scheduling algorithm has important practical significance.This paper will do some researches on chemical reaction optimization(Chemical Reaction Optimization)CRO algorithm for multi-objective task scheduling problem.CRO combined with the idea of genetic algorithm and simulated annealing algorithm,is a kind of swarm intelligence algorithm based on the interaction of molecules in the process of The CRO algorithm has been used to solve some optimization problems.The main achievements of this thesis can be summarized as follows:A multi-objective cloud task scheduling model for task completion time and cost is established.The task scheduling algorithm based on DAG considering the task priority constraint,the heterogeneity of the resources in the cloud environment,and the communication and priority constraint relations between tasks,can reflect the real characteristic of cloud environment and task.For multi-objective scheduling problem,a multi-objective chemical reaction algorithm is proposed,the four molecular collision criteria and molecular reactions in solving scheduling problem are redesigned,then it is used to solve the multi-objective cloud task scheduling model has been established.The simulation experiments show that the algorithm has good convergence,and achieved good results in solving multi-objective cloud task scheduling problem.Compared with the previous group of intelligent algorithms,it has strong ability of problem solving.Hereby,a hybrid algorithm based on particle swarm and chemical reaction optimization(HPSO-CRO)is proposed,taking advantage of the compensatory property of particle optimization algorithm and chemical reaction algorithm,the algorithm has good global search and local search ability.And it is used to solve the multi-objective task scheduling in cloud environment,finally the simulation experiments also indicate that the algorithm is feasible,it has good performance than particle optimization algorithm and chemical reaction algorithm in solving multi-objective task scheduling problems,it can spend less processing time and execution cost when the tasks are executed.