학술논문
Deep Reinforcement Learning Optimization Algorithm Designed for IRS-Assisted Edge Computing
Document Type
Conference
Author
Source
2023 IEEE 6th International Conference on Electronic Information and Communication Technology (ICEICT) Electronic Information and Communication Technology (ICEICT), 2023 IEEE 6th International Conference on. :835-840 Jul, 2023
Subject
Language
ISSN
2836-7782
Abstract
To reduce the computing delay of the terminals and improve the poor wireless channel environment, in this paper, a computing resource allocation algorithm for IRS-assisted edge computing is proposed. By introducing the intelligent reflecting surface (IRS) technique to reduce the edge computing latency, the computing offloading variables are optimized through an Alternating optimization approach, and a partial offloading strategy for MEC offloading is adopted. Deep reinforcement learning (DRL) is applied for coarse phase control of the IRS, and fine phase control of the IRS is achieved through fine phase optimization. In simulations, the proposed scheme is demonstrated to be effective and low latency.