MUMBAI, India, Aug. 12 -- Intellectual Property India has published a patent application (202621075910 A) filed by Symbiosis International Deemed University on June 18, 2026, for Dual Loop Adaptive Framework For Cloud Load Balancing Using Hierarchical Fuzzy Systems And Reinforcement Learning.
Inventors include Sukhada Bhoyar; and Dr. Bhupesh Kumar Dewangan.
The application for the patent was published on August 07, 2026, under issue no. 32/2026.
Abstract: ABSTRACT DUAL LOOP ADAPTIVE FRAMEWORK FOR CLOUD LOAD BALANCING USING HIERARCHICAL FUZZY SYSTEMS AND REINFORCEMENT LEARNING The present invention discloses a Dual Loop Adaptive Framework (DLAF) (100) for cloud load balancing that integrates a Hierarchical Fuzzy System (HFS) (120) with a Proximal Policy Optimization (PPO) based reinforcement learning meta controller (130). The fast loop HFS (120) decomposes four-dimensional server metrics comprising CPU utilization, memory pressure, queue length, and response latency into two parallel subsystems, a Computational Load Subsystem (121) and a Service Quality Subsystem (122), whose outputs are aggregated by a Meta Inference Layer (123) to produce server suitability scores at sub millisecond cadence. The slow loop PPO meta controller (130) periodically revises the HFS membership function parameters and rule weights based on accumulated QoS feedback through a compound reward function in the Reward Computation Engine (150). This dual loop architecture combines fuzzy logic interpretability with reinforcement learning adaptability, reducing rule base complexity by 88 percent while enabling continuous online optimization without manual threshold calibration. [
Disclaimer: Curated by HT Syndication.
