The Improvement Of Shuffled Frog Leaping Algorithm And Its Application To The Resource Scheduling Of Cloud Computing

As the mainstream of information technology nowadays,cloud computing can integrate the idle resource on the internet to form resource pool in virtue of virtualized technology.In order to providing better service,cloud computing would assign the related data and task dynamically based on the user demand.That is,it can split the complex task and application program into several fractions,thus forming giant system after the treatment of numerous servers.At last the obtained results are sent back to users or demanders after the calculation,search and analysis.In fact,the essence of cloud computing is the integration of idle resource and equipment on the internet,ultimately realizing real-time dynamic sharing.Then,how to realize an effective and efficient scheduling aiming at resource,equipment,data as well as application task becomes a significant topic in cloud computing fields.For the resource scheduling in the cloud computing environment,a majority of the researches are based on the following two methods: The first is that a certain amount of task would be assigned to the equipments that have some vacant resource.The second is that some resources can be preferentiallyassigned when they require much task.However,these dispatching schemes are too simple,whether in finishing time or loading content,to meet the service requirement of current clients.Above all,this work is meant to design a resource scheduling scheme to ensure both of the speed and loading balance.This work improves the shuffled frog leaping algorithm(SFLA)and applied it to the resource scheduling in the cloud computing environment.Aiming at some existing issues of SFLA in resource scheduling of cloud computing,such as the randomicity at the initial state,blindness of local search,the unicity in step size updating,low convergence speed,easily to be locally optimum,lacking intergroup exchange in shuffle and so on,the following three aspects of wok has been done to improve the algorithm.First of all,initialize the population by integrating the SY-MM algorithm and randomly generation mode,thus producing the frog population with nice fitness and diversity.Then,it can realize a self-updating of step size for frog individual,promote the local research ability,fasten the convergence speed of algorithm and improve the local search accuracy by some proposed methods,such as improving the step size formula,introducing weight into frog individual.Finally,it can promote the information exchange in frog population by exchanging the chromosome using genetic algorithm and single-point-order exchange for chromosome,obtaining the offspring individual with better adaptation and avoiding the production of local optimum.In order to verify the feasibility and validly of the proposed algorithmapplied at resource scheduling in the cloud computing environment,this work conducts the simulation experiment by means of CloudSim simulation platform.Moreover,the initialization degree,time and loading between the proposed algorithm and initial frog leaping algorithm,genetic algorithm as well as particle swarm optimization algorithm has been also compared and analyzed.The experimental results demonstrate that the proposed algorithm possesses excellent property,which can shorten the completion time of resource scheduling in the cloud computing environment and balance the loading content of every resource.