MUMBAI, India, Aug. 31 -- Intellectual Property India has published a patent application (202621082003 A) filed by Symbiosis International Deemed University on July 02, 2026, for Machine Learning-Based Deadline-Aware Adaptive Task Scheduling For Dynamic Cloud Resource Allocation.
Inventors include Madhura Hawelikar; and Dr. Bhupesh Kumar Dewangan.
The application for the patent was published on August 28, 2026, under issue no. 35/2026.
Abstract: ABSTRACT MACHINE LEARNING-BASED DEADLINE-AWARE ADAPTIVE TASK SCHEDULING FOR DYNAMIC CLOUD RESOURCE ALLOCATION The present invention discloses a predictive and adaptive deadline-aware task scheduling system (100) for dynamic cloud resource allocation using machine learning-based workload forecasting. The system (100) comprises a workload data acquisition module (110), a machine learning prediction module (120) that generates predicted future CPU load values using linear regression trained on historical time-series workload data with engineered lag features, a prediction output interface (130), a prediction service module (140), a predictive adaptive scheduler (150) incorporating a load threshold evaluator (152), a VM scaling controller (154), a deadline-aware task sorter (156), and a least-loaded VM assigner (158), a cloud simulation environment (160), and a performance metrics evaluation module (170). The system proactively adjusts virtual machine allocation based on predicted workload thresholds, prioritizes deadline-critical tasks during high load periods, and distributes tasks to least- loaded virtual machines, achieving significant reductions in makespan and SLA violation rate compared to conventional reactive scheduling algorithms. [
Disclaimer: Curated by HT Syndication.
