MUMBAI, India, July 13 -- Intellectual Property India has published a patent application (202621067635 A) filed by Parul University Parul Institute Of Engineering And Technology on May 29, 2026, for Cognitive Cluster-Based Task Scheduling System For Heterogeneous Cloud Computing Environments.
Inventors include Divya. R; and Swapnil M Parikh.
The application for the patent was published on July 10, 2026, under issue no. 28/2026.
Abstract: A cloud scheduling system integrates two-stage hybrid clustering with deep reinforcement learning to optimize heterogeneous cloudlet allocation to virtual machines. The system employs density-based clustering using Hierarchical Density-Based Spatial Clustering of Applications with Noise (HDBSCAN) to automatically group cloudlets without predefined cluster counts and filter noise points, followed by probabilistic refinement using Gaussian Mixture Model (GMM) to improve cluster boundary precision for cloudlets with overlapping resource characteristics. A Dueling Double Deep Q-Network (D3QN) scheduler evaluates system states including CPU utilization, memory usage, queue length, cluster size, and workload distribution to select optimal virtual machine allocation. A multi- objective reward function minimizes makespan and energy consumption while maximizing load balancing and resource utilization. A monitoring and feedback unit provides continuous performance tracking to update the reinforcement learning policy, enabling adaptive improvement over time. The system achieves enhanced scalability, adaptability, reduced execution delay, balanced load distribution, and improved overall performance in heterogeneous cloud computing environments .
Disclaimer: Curated by HT Syndication.
