Hybrid Deep Q-Network-Based Multilevel Access Controlled Task Scheduling and Load Balancing in Cloud Environment

Abstract

Recently, cloud computing has emerged as a promising technology, which enables service providers to deliver computing resources and storage service to organizations and individuals across the internet. Efficient task scheduling and load balancing are the critical challenges in dynamic cloud environments that affects resource utilization, response time, and energy efficiency. However, conventional models often struggle with long-term adaptability, security assurance, and multilevel access control. To address these issues, this manuscript proposes a Hybrid Deep Q-Network with multilevel authentication (HDQN-MLA) algorithm for multilevel access controlled task scheduling and load balancing in a cloud environment. The proposed model integrates Boltzmann exploration and modified $\varepsilon$-greedy to optimize exploration and exploitation, while a multilevel authentication mechanism ensures secure and controlled cloud access. The experimental evaluation of the proposed task loads demonstrates that it achieves superior throughput, less response time, reduced makespan, and lower energy consumption, which is better compared to the existing deep reinforcement learning and metaheuristic-based models. The results of the proposed model indicate effectiveness in achieving secure, efficient, and adaptive task scheduling in cloud environment.