Optimizing Dynamic Computing Environments Using Adaptive Intelligent Load Balancing

Cloud computing environments are inherently dynamic, characterized by fluctuating workloads and resource availability, which poses significant challenges for traditional load balancing algorithms. Static or rule-based approaches often fail to adapt effectively, leading to inefficient resource utilization and degraded Quality of Service. This paper presents an adaptive and intelligent load balancing system leveraging Deep Reinforcement Learning (DRL). We developed a framework integrating a DRL agent, implemented in Python, with the CloudSim Plus simulator using a Py4J gateway. The DRL agent learns optimal task assignment policies through interaction with the simulated cloud environment, aiming to minimize task execution times and improve overall resource utilization. This work explores the potential of DRL to autonomously manage resources in complex cloud settings, offering a pathway towards more efficient and responsive cloud service delivery compared to conventional load balancing strategies.