Hybrid Metaheuristic-Based Cloud Task Scheduling Using Genetic Algorithm and Salp Swarm Algorithm

Today, cloud computing has made major changes and improvements to storing, processing, and managing data using already-developed models such as IaaS, PaaS, and SaaS. Task scheduling in cloud computing is a process by which a task is assigned and processed dynamically across the cloud resources such as Virtual Machines, VMs, containers, or physical servers wherever available based on some constraints such as availability, priority, time constraint, and resource utilization. Moreover, the task scheduling problem in cloud environments is a core challenge in defining a system's efficiency and overall cost. In this context, the more classic approach of scheduling algorithms, such as FCFS and RR, do not properly manage complex and heterogeneous cloud environments. To overcome these challenges, this paper employs metaheuristic algorithms, which are particularly suited for cloud task scheduling due to their ability to handle NP-hard optimization problems. Metaheuristic methods are near-optimal techniques for searching vast and complex solution spaces, simultaneously exploring and exploiting the solution space. This paper presents a hybrid approach by combining the Genetic Algorithm (GA) and Salp Swarm Algorithm (SSA) to identify an efficient approach to task scheduling in cloud computing; by combining these two techniques, the proposed hybrid model effectively balances exploration and exploitation, leading to improved task distribution and execution time optimization. The experiments were performed using 10 tasks of different lengths on 5 Virtual Machines with randomly generated processing capabilities. The outcome of the evaluation proves that the proposed combination of HGA and SSA surpasses individual GA and SSA in distributing the tasks while minimizing the maximum execution time to 115.60 sec, with an optimization time of 0.18 sec. The work emphasizes using metaheuristic algorithms to optimize scheduling methods and proposes useful data for cloud computing implementations.