Research And Implementation Of Improved Differential Evolution Evolution Algorithm Based On Cloud Computing

The existing algorithms of cloud computing scheduling class are mostly aimed at the problems of combination class optimization NP(Non-Deterministic Polynomial Problems),such as FIFO(First In First Out),fair scheduling,capacity scheduling and so on.This type of scheduling algorithm is difficult to fully meet the practical application,because they all have their own irreparable defects,and slowly some new algorithms appear on this,the new algorithm on the one hand to improve some of the defects of the traditional algorithm,but like Genetic algorithms,particle swarm optimization,ant colony optimization and other optimization algorithms still have shortcomings,and further optimization is needed.The Differential Evolution Algorithm(DE)was introduced in this context.It is a highly efficient and global algorithm.The differential evolution algorithm is based on the group and the group-based heuristic.The search algorithm,any individual in the group is a feasible solution.In the course of population evolution,DE algorithm will undergo mutation,crossover,and selection operations in sequence.This is similar to genetic algorithm,but the difference is that the definitions of mutation operation,cross-operation,and selection operation are different;DE algorithm is on the other hand.See also a simulated biological evolutionary algorithm,which will continue to evolve and iterate,and finally retains the fine individuals that can adapt to the environment.What is more dominant than the genetic algorithm is that the DE algorithm is a population-based heuristic search algorithm,which uses real numbers to encode,and uses a one-to-one survival strategy and relatively simple mutation mutations to reduce genetic complexity.To solve the optimization problem in a complex environment;so far,differential evolution algorithm has achieved good results in applications such as signal processing,food safety,and robotics.For this topic,the research direction mainly starts with the basic theory of DE algorithm,first analyzes its entire basic flow,and sets the constraints on its various important parameters,such as the initial population individual number NP,individual Dimension D,maximum population iterations G,current iteration times t,variation factor F,crossover factor CR,fitness choice,etc.Secondly,a lot of experiments are done on DE algorithm to find out the disadvantages of DE algorithm: when the population size is large,the population convergence is very slow;when the population size is small,local optimal individuals are easily obtained,and a new improved algorithm is proposed according to the disadvantages.: Change the range of variation factors and crossover factors,and propose a separate differential evolution algorithm.The experimental results show that the improved separation evolution algorithm plays a significant role in solving the problem.This topic aims to verify and improve the DE algorithm,which will be introduced into the cloud computing simulation environment CloudSim,in order to simulate the cloud computing task scheduling process,and Finally through experiments to verify the performance of the improved algorithm.Prove the role of the improved algorithm in cloud computing task scheduling.