Multi-Objective dependent Tasks Scheduling using AWGWO algorithm in Cloud Computing

Abstract

This paper introduces a method to solve workflow scheduling algorithm problems in cloud computing. The workflow scheduling is an appropriate scheduling algorithm to reduce the total execution time (TET) and total execution cost (TEC) of tasks when tasks are interdependent. The nature of workflow scheduling is np- hard, so traditional algorithms have failed to solve the workflow scheduling problem. This paper proposes a meta-heuristic algorithm named the AWGWO (adaptive weights grey wolf optimization) algorithm. The AWGWO is a new variant of GWO. This new variant is modified in two ways: first, considering the new wolf leader's name, gamma, and second, the weights assigned to wolf leaders are adaptive instead of the average weight of wolf leaders, similar to the original GWO algorithm. This modification of GWO maintains a balance between exploration and exploitation because the early phase describes exploration, while the latter phase describes exploitation. This balance enables efficient assignment of the tasks to virtual machines and reduces the TET and TEC of workflows. The proposed algorithm outperforms the ACO and GWO algorithms.