IKH-EFT: An improved method of workflow scheduling using the krill herd algorithm in the fog-cloud environment

Abstract

Given the increase diversity of smart devices and objectives of the application management such as energy consumption, makespan users expect their requests to be responded to in an appropriate computation environment as properly as possible. In this paper, a method of workflow scheduling based on the fog-cloud architecture has been designed given the high processing capability of the cloud and the close communication between the user and the fog computing node, which reduces delay in response. We also seek to minimize consumption and reduce energy use and monetary cost in order to maximize customer satisfaction with proper scheduling. Given the large number of variables that are used in workflow scheduling and the optimization of contradictory objectives, the problem is NP-hard, and the multi-objective metaheuristic krill herd algorithm is used to solve it. The initial population is generated in a smart fashion to allow fast convergence of the algorithm. For allocation of tasks to the available fog-cloud resources, the EFT (earliest finish time) technique is used, and resource voltage and frequency are assumed to be dynamic to reduce energy use. A comprehensive simulation has been made for assessment of the proposed method in different scenarios with various values of CCR. The simulation results indicate that makespan exhibits improvements by 9.9, 8.7% and 6.7% on average compared with respect to the methods of IHEFT, HEFT and IWO-CA, respectively. Moreover, the monetary cost of the method and energy use have simultaneously decreased in the fog-cloud environment.