Efficient task scheduling in cloud computing is crucial for managing dynamic workloads while balancing performance, energy efficiency, and operational costs. This paper introduces a novel Reinforcement Learning-Driven Multi-Objective Task Scheduling (RL-MOTS) framework that leverages a Deep Q-Network (DQN) to dynamically allocate tasks across virtual machines. By integrating multi-objective optimization, RL-MOTS simultaneously minimizes energy consumption, reduces costs, and ensures Quality of Service (QoS) under varying workload conditions. The framework employs a reward function that adapts to real-time resource utilization, task deadlines, and energy metrics, enabling robust performance in heterogeneous cloud environments. Evaluations conducted using a simulated cloud platform demonstrate that RL-MOTS achieves up to 27% reduction in energy consumption and 18% improvement in cost efficiency compared to state-of-the-art heuristic and metaheuristic methods, while meeting stringent deadline constraints. Its adaptability to hybrid cloud-edge architectures makes RL-MOTS a forward-looking solution for next-generation distributed computing systems.
Yu et al. (Wed,) studied this question.