A hybrid evolutionary algorithm to improve task scheduling and load balancing in fog computing

This paper introduces a hybrid evolutionary task scheduling and VM placement algorithm (HETSVP) designed for dependable fog computing task scheduling and VM placement. We address the optimization of task execution time and resource balance concurrently by integrating an improved particle swarm optimization algorithm with a new VM placement strategy. We utilize a direct binary encoding technique, where the particle's location information is expressed using 0 and 1, and particle velocity ranges within [0, 1]. In the context of discrete particle swarm, each particle's position will signify a potential task scheduling plan. Also we provide the adaptive contraction factor, which improves the efficiency of the particle swarm optimization approach. On the other hand, due to the importance of VM placing we introduce a new placement strategy that according to the available of PMs resources, the VM placement operation is performed. We then conduct simulation experiments in ifogsim environment to assess the performance of HETSVP. The results demonstrate that compared to the MinMin and MaxMin algorithm, HETSVP reduces makespan by 11% and enhances energy consumption by 15%. Additionally, compared to Genetic and ACO algorithms, HETSVP achieves 13%, 5% reduction in makespan, and improves energy consumption by 12%, 5%. Moreover, the simulations indicate that the proposed method can improve the fog environment's reliability. Furthermore, compared to other methods, HETSVP exhibits a better degree of imbalance and makespan results.