Energy Efficient Task Scheduling in Serverless Computing Using PSO and ACO Algorithms

Serverless computing offers significant scalability and cost efficiency but faces challenges in optimizing task scheduling to ensure energy efficiency and Service Level Agreement (SLA) compliance. This research proposes a hybrid task scheduling framework combining Particle Swarm Optimization (PSO) and Ant Colony Optimization (ACO) to address these challenges in Function as a Service (FaaS) environments. The PSO algorithm handles global task distribution, while ACO refines local scheduling decisions through adaptive pheromone updates. The approach utilizes real-time metrics such as latency, resource load, and SLA deadlines to schedule tasks efficiently. Experimental results show that the PSO-ACO framework reduces energy consumption by up to 30%, improves function latency, and achieves over 91% SLA compliance across multiple workload scenarios. When compared to traditional algorithms like FCFS, RR, and EDF, the proposed method outperforms in terms of energy savings and SLA adherence. Additionally, the model demonstrates robustness under burst workloads, enhancing both Energy Delay Product (EDP) and Normalized Power Savings (NPS). The results suggest that the hybrid PSO-ACO approach offers a scalable, energy efficient solution for task scheduling in serverless systems. This work contributes to the development of more sustainable cloud computing platforms, with practical applications in real-time services, IoT, and smart city applications, where energy efficiency and SLA compliance are crucial.