Time-oriented Cloud Workflow Scheduling Optimization Based On Genetic Algorithm

Cloud computing is inherited and developed on the basis of decades of unremitting efforts in the fields of virtualization,distributed computing,utility computing,and grid computing.At present,although the global cloud computing industry is in its infancy,it is developing rapidly and has a promising future.The workflow in the cloud computing environment,that is,the "cloud workflow" refers to the application of workflow related technologies to the cloud computing domain to improve the service quality of cloud computing and realize the rational allocation of cloud computing resources.Cloud workflow scheduling refers to mapping/distributing tasks in a workflow to appropriate cloud computing resources with an optimized method while satisfying workflow timing constraints,system performance,and customer needs.The efficiency of cloud workflow scheduling is an important indicator to measure the performance of cloud workflow system.Effective and reasonable cloud workflow scheduling can not only improve the overall performance of cloud workflow system,but also improve the performance of other components in cloud workflow system.With the continuous development of cloud computing and the expansion of application fields,cloud workflow scheduling has become an important research topic.For the cloud workflow scheduling optimization problem,this paper introduces the related concepts of cloud workflow,and establishes a cloud workflow model for scheduling optimization and various time calculation methods in cloud workflow environment.Cloud workflow scheduling is an NP-hard problem,and related literature indicates that genetic algorithms are more suitable for solving such problems.However,some current genetic algorithms have certain limitations,either the search space is incomplete or the search efficiency is too low.Aiming at the characteristics of cloud workflow scheduling optimization problem and some limitations of current genetic algorithm,this paper proposes RKLGA,which is a real number encoding based on random keys.Any scheduling scheme has a corresponding encoding scheme to ensure the search space.Completeness.Hierarchical techniques have been added to population initialization,genetic manipulation,and decoding operations to improve search efficiency and population diversity.For the optimal individual in the population,a virtual machine load balancing improvement strategy is adopted to enhance the field optimization ability of the chromosome.In order to verify the validity of the proposed RKLGA,four typical cloud workflow cases are compared,and the algorithm is implemented based on C++program.Three typical algorithms are selected for comparison experiments.The experimental results show that compared with other algorithms,RKLGA not only has high search efficiency,but also can quickly find the optimal solution,which can be widely applied to various workflows.