Multi-objective Cloud Workflow Scheduling Strategy Based On Evolutionary Algorithm

Workflow is a common model for providing scientific experiments that consist of many tasks,data flows,and computational dependencies.The process of mapping workflow tasks to computing resources(VMs)for execution(reserving dependencies between tasks)is referred to as workflow scheduling.Workflow technology is a general-purpose model that describes a wide range of scientific applications in distributed systems.It plays an important role in many fundamental sciences such as physics,chemistry,biology,and computer science.Cloud computing provides a promising platform for executing large applications with large computing resources,available on-demand in cloud models,where users charge based on their resource usage and required quality of service(QoS)specifications.Although there are many existing workflow scheduling algorithms in traditional distributed or heterogeneous computing environments,it is difficult to apply directly because cloud and service-based resource management methods and pay-per-use pricing strategies are different from traditional heterogeneous environments.In the cloud environment.We propose an evolutionary multi-objective optimization(IEMS)based algorithm to solve the workflow scheduling problem on the Infrastructure as a Service(IaaS)platform.The algorithm uses new coding methods and genetic operators for multi-target cloud workflow scheduling,and improves the individuals of the initial population to speed up the search.Extensive experiments on real-world workflows and randomly generated workflows show that,in most cases,our algorithms can achieve better solutions than existing QoS-optimized scheduling algorithms.The experiments conducted were based on Amazon EC2's on-demand instance types;however,the algorithm easily scaled to the resources and pricing models of other IaaS services.