MUMBAI, India, Oct. 12 -- Intellectual Property India has published a patent application (202641118552 A) filed by Dr. B. Lakshmi Bhavani; Dr. N. Vimala; B Prasanthi; Dr. Ravi Kumar Athota; Dr. D. Anupama; Dr. K Shravani; A. Lakshmi Bhargav; and Ms. Priyanka Chaumwal on October 04, 2026, for Machine Learning-Based Predictive Edge-Computing System For Intelligent Resource Allocation And Dynamic Workload Optimization.
Inventors include Dr. B. Lakshmi Bhavani; Dr. N. Vimala; B Prasanthi; Dr. Ravi Kumar Athota; Dr. D. Anupama; Dr. K Shravani; A. Lakshmi Bhargav; and Ms. Priyanka Chaumwal.
The application for the patent was published on October 09, 2026, under issue no. 41/2026.
Abstract: [052] A machine learning-based predictive edge-computing system is disclosed for intelligent resource allocation and dynamic workload optimization across heterogeneous edge nodes. The system acquires time-dependent workload, resource and network telemetry and constructs a synchronized temporal feature state. A machine-learning forecasting engine predicts future workload-demand and resource-availability states, while a confidence and drift evaluator determines reliability of the predicted states. A candidate allocation generator constructs technically feasible placement, scaling, scheduling and migration actions, and a multi-objective optimizer selects an action using predicted service performance, resource headroom, energy or network cost, transition overhead and prediction confidence. An execution and migration controller applies the selected allocation and obtains measured outcomes. Prediction residuals and action-performance residuals are returned through a closed feedback path to adapt subsequent forecasting and optimization. The architecture enables anticipatory allocation while suppressing unnecessary reconfiguration caused by uncertain predictions, thereby improving edge-resource utilization, service latency, energy efficiency and execution stability. Accompanied Drawing [FIGS. 1-2].
Disclaimer: Curated by HT Syndication.
