Research On Multi-objective Scientific Workflow Scheduling Algorithm Under Cloud Environment

Along with the fast development of cloud computer technology and wide usage of scientific workflow application in cloud computer,the scientific workflow scheduling problem attracts common attention,and how to satisfy the limits on time,energy consumption and expenses becomes the hot focus in the research.During the multi-scientific workflows scheduling algorithm study,now the security of some algorithms is still not improved,then lack of security and integrity.Meanwhile,the multi-objective scientific workflows scheduling based on particle swarm optimization(PSO)exists deficiency of partial optimal solution in the convergence later stage,so the performance still require enhancement.About the above mentioned two problems,the article separately brings out the scheduling algorithm of multi-scientific workflows security limit and multi-objective scientific workflows scheduling algorithm based on PSO.The innovation point and main work of this article embody in the following sides:1.According to the partial unimproved safety of multi-scientific workflows scheduling algorithm under current cloud environment,it advises multi-scientific workflows security-deadline constraint cost optimization algorithm(MSW-SDCOA).First,MSW-SDCOA can compress scientific workflows based on data depending relationship then decrease task node quantity and save scheduling cost;Then,it also can form scheduling series by improving HEFT(heterogeneous earliest-finish-time)algorithm to realize overall multi-objective optimization scheduling;Finally,through the information update strategy and heuristic information in ant colony optimization(ACO),to let the evolution of ant colony more objective then improve the cost optimization effective further.2.As to the problem of deficiency of partial optimal solution in the convergence later stage of multi-objective scientific workflows scheduling based on PSO,this article proposes the multi-objective scheduling algorithm for scientific workflow based on immune particle swarm optimization(IPSO-SWS),which combines the immune algorithm(IA)introduced with PSO,then utilizes the concentration self-adjustment system of IA to guarantee the diversification of particle(antibody)then keep that of population,it also sets antibody memory stock to let particle(antibody)evolve to the correct direction then finds the overall optimized scheduling fast.3.This article imitates the scheduling of scientific workflows by using CloudSim(a simulated platform),then realizes the MSW-SDCOA and IPSO-SWS algorithm.After multiple groups experiment analysis,it verifies the effectiveness of algorithms mentioned in this article from finish time,cost and cloud resource availability.The MSW-SDCOA algorithm is 14%higher than MW-DBS on expenses optimization effect after contrast test,while the IPSO-SWS algorithm is 8.4%lower than CDCGA on running time,but 8%higher than that on expenses optimization.In conclusion,this article makes research for the two problems required to solve in current scientific work-flow scheduling algorithm,also proposes the corresponding solution algorithms.The study in this article doesn't only guarantee the user safety and time limit,but also decrease the task implement cost of scientific work-flow under cloud environment,then reduces the energy consumption of cloud-computer and promotes the development of energy conservation and emission reduction in our country.Therefore,the study in this article is not only innovate in theory but also meaningful in actual applications.