Task scheduling using memetic intelligent water drops algorithm based on tabu search: a case study on azure workflows

Task scheduling is one of the NP-Complete optimization problems playing a vital role in distributed systems, especially cloud computing and fog computing. The delay due to scheduling affects the pay-as-you-go model negatively and imposes excess costs on users. In this research, a new meta-heuristic based on intelligent water drops algorithm is proposed which uses the tabu search method as a local search. The ideal ability of the intelligent water drops algorithm in general search, along with the remarkable capacity of tabu search in local search guarantees the balance between exploration and exploitation processes in evolutionary cycles. Subsequently, several experiments were performed to evaluate the proposed method on standard benchmark functions, random DAGs, and actual workflows derived from the Microsoft Azure platform, respectively. The results demonstrated that the proposed method also has superiority over the state-of-the-art technologies in terms of algorithmic criteria such as convergence rate and running time, as well as scheduling criteria such as waiting time and makespan.