Trust Aware Multi-Objective Metaheuristics for Workflow Scheduling in Cloud Computing

Nowadays, in the world of modern businesses, it has become a trend to switch to cloud computing by deploying their workflow applications. Nevertheless, workflow scheduling is a challenging task because of the dependency among its tasks, high computational intensity, and heterogeneous behavior of resources. Execution time and trust for virtual machines on which the tasks are to be scheduled is the most difficult job during resource allocation and raises considerable problems. Thus, management of reducing of Makespan time and allocation of tasks to the virtual machine having the highest trust value at the same time is necessary. It becomes a multi-objective problem. Regarding the current issue, we address a multi-objective function of trust and the execution time for the selection of virtual machines for the allocation of workflow tasks. We have proposed MOGAACO, a multi-objective function-based approach using Genetic Algorithm (GA) and Ant Colony Optimization (ACO) for scheduling workflows. For the generation of initial pheromone, ACO invokes GA. With the designed multi-objective function three parameters are evaluated, Load Balance Level (LBL), Makespan, Resource Utilization (RU). CloudSim simulator is used for simulation of experimental results and the proposed algorithms perform better than the basic ACO, GA, and heuristic algorithm MIN-MIN.