MUMBAI, India, Aug. 31 -- Intellectual Property India has published a patent application (202621081986 A) filed by Symbiosis International Deemed University on July 02, 2026, for Real-Time Adaptive System And Method For Cpu Stress Detection And Automated Stabilization Using Machine Learning.
Inventors include Dr. Shreyas Rajendra Hole; and Omkar Pant.
The application for the patent was published on August 28, 2026, under issue no. 35/2026.
Abstract: ABSTRACT REAL-TIME ADAPTIVE SYSTEM AND METHOD FOR CPU STRESS DETECTION AND AUTOMATED STABILIZATION USING MACHINE LEARNING The present invention discloses a real-time adaptive system (100) and method for CPU stress detection and automated stabilization on a standalone computing device. The system (100) comprises a monitoring module (110) for continuously polling CPU and memory metrics, a Health Tracking Score engine (120) for computing a composite score quantifying deviation from a performance baseline, an anomaly detection module (130) employing a trained Isolation Forest machine learning model for identifying anomalous process behavior, a decision engine (140) for evaluating detected conditions and selecting corrective actions, an auto-fix engine (150) implementing three escalating intervention levels comprising process priority adjustment (L1), CPU core affinity restriction (L2), and process suspension (L3), a logging and learning module (160) for recording action effectiveness and updating decision logic, and a user interface dashboard (170) for real-time visualization. The system operates entirely in user-space without kernel modifications and progressively improves stabilization response through feedback-driven learning. [
Disclaimer: Curated by HT Syndication.
