A Multi-Level Feedback Queue Optimization Method Based on Reinforcement Learning and Dynamic Time Slice Scheduling
Year: 2025
Type: article
Abstract: This invention pertains to the field of reinforcement learning and discloses a multi-level feedback queue optimization method [2] based on reinforcement learning and dynamic time slot scheduling. It involves separately obtaining process feature data, system load data, user experience data, and hardware-related data. The acquired data is preprocessed to ensure accuracy and reliability. Based on the preprocessed data, the data sets for process features, system load, user experience, and hardware-related are formed through numbering. Using these numbered datasets, the process feature index, system load rate, user experience index, and hardware load rate are calculated. A comprehensive evaluation of queue priority values is conducted, and multi-level feedback queue optimization is implemented based on these evaluated queue priority values. This approach considers multiple dimensions of scheduling methods, enhancing the quality and accuracy of scheduling decisions, avoiding certain processes from occupying too many resources and affecting the operation of other processes, making multi-level feedback queue optimization more equitable.
Source: Automation and Machine Learning
Author: Zhong Wenxuan
Institution: Hangzhou Dianzi University
Language: English
Topic: Real-Time Systems Scheduling
Subfield: Hardware and Architecture
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W4411414254