Deadline‐constrained coevolutionary genetic algorithm forrnscientific workflow scheduling in cloud computing

摘要

The cloud infrastructures provide a suitable environment for the execution of large‐scale scientificrnworkflow application. However, it raises new challenges to efficiently allocate resources for thernworkflow application and also to meet the user's quality of service requirements. In this paper,rnwe propose an adaptive penalty function for the strict constraints compared with other geneticrnalgorithms. Moreover, the coevolution approach is used to adjust the crossover and mutationrnprobability, which is able to accelerate the convergence and prevent the prematurity. Wernalso compare our algorithm with baselines such as Random, particle swarm optimization,rnHeterogeneous Earliest Finish Time, and genetic algorithm in a WorkflowSim simulator on 4rnrepresentative scientific workflows. The results show that it performs better than the otherrnstate‐of‐the‐art algorithms in the criterion of both the deadline‐constraint meeting probabilityrnand the total execution cost.