Research On Resource Allocation Based On Ant Colony Optimization Algorithm Under Cloud Computing

As Internet technology and information technology develop in an explosive way,more and more enterprises and individuals have tended to carry out business and work using the Internet,leading to continuous accumulation of information on the Internet.Along with the connection of mass mobile devices to the Internet,the quantity of information has increased more and more rapidly.And the traditional computing mode has been unable to meet user demand for the Internet.Under this general background,an emerging distributed computation-based computation model-“cloud computing”-has emerged in response to the needs of the times.In the developing process of cloud computing,the solving of resource allocation problem,a key technology of cloud computing,is still in the stage of research and development.Resource allocation in cloud environment aims mainly to solve two problems: the one is about full use of the resources in cloud environment to set up an effective operating mechanism by coordinating the balance between performance and load;the other is about whether a resource allocation strategy can effectively adjust the damage,loss or changed demand of a certain resource in cloud environment in real time.This paper aims primarily to study the optimization of resource allocation algorithm under cloud computing.On the basis of summing up the previous work,this paper made the following studies and innovative points:1.It described the principle of ant colony optimization and analyzed the advantages and disadvantages of ant colony optimization,and on this basis put forward some suggestions for the improvement of ant colony optimization: crossed and mutated pheromone concentration.Mutation can effectively prevent the algorithm from falling into the local optimum,and a new solution obtained from crossover and mutation is usually closer to the global optimal solution than the previous two good solutions,which can help speed up the convergence of solution and increase the accuracy.The use of this iterative optimal solution to enhance the concentration of pheromone on this path can increase the accuracy of algorithm convergence.Anexperiment verifies that the performance of the improved ant colony optimization indeed has been improved somewhat.2.Due to the uniform distribution of pheromone concentration in the initial operating stage of ant colony optimization,a blind search wasinevitably conducted,which had an impact on the convergence rate of the whole algorithm.On this basis,RAAG(Refinement Algorithm for ACO and GA),a new algorithm was created from the fusion of genetic algorithm and ant colony optimization,and the concrete realization of the fusion was described.The application of genetic algorithm’s global searching capability in prophase in this algorithm could make up for the shortage of ant colony optimization’s searching ability in prophase.With the optimal solution of genetic algorithm transformed into the initial pheromone of ant colony optimization,the performance of the whole algorithm was improved.The performance and feasibility of this algorithm were verified through CloudSim simulation experiment,the work in this paper was summarized and a research on the problem of resource allocation under cloud computing was expected in the end.