Research On Workflow Scheduling Based On Improved Particle Swarm Algorithm

In recent years,due to the rapid development of cloud computing technology,the business model of on-demand payment and personalized computing requirements pose a huge challenge to the existing computing resources in workflow scheduling.Therefore,researchers are committed to introducing meta-heuristic scheduling algorithms to solve the problem of resource utilization.Most of them focus on the load balancing supply of tasks to generate efficient resource utilization.However,this focus will increase the execution time of large-scale tasks.As a result,the scheduling efficiency of large-scale tasks is too low.In addition,most scheduling algorithms on the market currently only perform scheduling for the problem of singularity,ignoring comprehensive considerations.The particle swarm optimization(PSO)algorithm,as a popular algorithm used in workflow scheduling,has the problems of premature convergence and low population diversity,in addition,the existing particle swarm algorithm on the market has a single population and is prone to fall into a local optimal solution,and cannot get the final optimization deployment plan.To address these problems,this paper proposes a new multi-objective cloud workflow scheduling based on an improved particle swarm optimization algorithm.The main contributions of this article are as follows:(1)This paper proposes a weighted multi-objective evaluation model,that is,based on the traditional model,it additionally considers the processor frequency characteristics,processor performance differences and the concept of "more work for those who can work",and constructs a model that includes workflow tasks.A comprehensive evaluation equation integrating the execution cost equation,workflow task execution time equation and balancing cluster load equation(multiobjective evaluation model including weighted features).(2)This paper proposes a new particle swarm optimization algorithm(IPSO-SA)based on simulated annealing: it uses multi-objective particle swarms instead of single-objective particle swarms to expand the range of optimal task deployment schemes for particle search,improve the diversity of the population and reduces the probability of premature convergence.In addition,each particle in the particle swarm uses the improved Metropolis criterion to update the particle position,which can effectively improve the search range of particles in the early iteration and accelerate the search range of convergent particles in the later iteration.(3)This paper compares and analyzes the proposed algorithm with the existing improved algorithms of related particle swarms(hybrid GA-PSO algorithm and IMPSO algorithm).The final experimental data show that by using the multi-objective evaluation model including weighted features constructed in this paper,the comprehensive performance of IPSO-SA is improved by 10.1%and 6.2% compared with the hybrid GA-PSO algorithm and the IMPSO algorithm,respectively.