An Integrated Grey Wolf Optimizer with Nelder-Mead Method for Workflow Scheduling Problem

Abstract

Cloud computing is one of the latest distributed system paradigms that comes with the opportunity of running workflows at reduced costs since it does not require owning any infrastructure Scientific workflows refer to a series of computations that facilitates data analysis in both structured & distributed manners. This paper formulated a new mathematical modeling for scientific workflow scheduling (SWS) problem. The formulated optimization problem is considered a multiobjective optimization task where MakeSpan, Cost, Energy, and FlowTime are handled as the objective functions. This study proposes a new hybrid optimization algorithm based on Grey Wolf Optimizer and Nealder Mead Method for solving multi-objective SWS problems. The obtained results based on several workflow templates showed that the proposed algorithm outperformed the well-known Heterogeneous Earliest First Time (HEFT) and Distributed HEFT (DHEFT). Moreover, its performance was better than that of the benchmarking algorithms.