Decision variable contribution based adaptive mechanism for evolutionary multi-objective cloud workflow scheduling

Abstract Workflow scheduling is vital to simultaneously minimize execution cost and makespan for cloud platforms since data dependencies among large-scale workflow tasks problem involve interactive decision variables. So far, the cooperative coevolution approach poses competitive superiority in resolving problems by transforming original into a series of small-scale subproblems. However, static transformation mechanisms cannot separate variables, whereas random encounter low efficiency. To tackle these issues, this paper suggests decision-variable-contribution-based adaptive evolutionary (VCAES short). be specific, VCAES includes new estimation method quantify contribution each variable population advancement terms both convergence diversity, dynamically classifies variables according their contributions during previous iterations. Moreover, mechanism adaptively allocate evolution opportunities constructed group Thus, with strong impact on are assigned more accelerate approximate Pareto-optimal fronts. verify effectiveness proposed VCAES, we carry out extensive numerical experiments real-world workflows compare it four representative algorithms. The results demonstrate problems.