RIS-Aided NOMA Downlink Networks: A DRL-Based Approach

Waqas Khalid; Heejung Yu

RIS-Aided NOMA Downlink Networks: A DRL-Based Approach

Department of Electrical and Electronic Engineering

Research output: Chapter in Book/Conference proceeding › Conference contribution › peer-review

Abstract

This paper addresses the power allocation problem in a reconfigurable intelligent surface (RIS)- aided non-orthogonal multiple access (NOMA) downlink system for short packet communications (SPC). A deep reinforcement learning (DRL) framework based on the proximal policy optimization (PPO) algorithm is proposed to jointly optimize transmission power and power allocation coefficients, maximizing achievable data rates under power constraints. The DRL-based approach offers adaptability, low training overhead, and efficiency in dynamic environments.

Original language	English
Title of host publication	35th Joint Conference on Communications and Information, Sokcho, South Korea
Publication status	Published - 23 Apr 2025

Keywords

Reconfigurable intelligent surface (RIS)
Short packet communications (SPC)
Deep reinforcement learning (DRL)

Cite this

@inproceedings{aa9c3acbbd3f463b8e4974b30b25f8e7,

title = "RIS-Aided NOMA Downlink Networks: A DRL-Based Approach",

abstract = "This paper addresses the power allocation problem in a reconfigurable intelligent surface (RIS)- aided non-orthogonal multiple access (NOMA) downlink system for short packet communications (SPC). A deep reinforcement learning (DRL) framework based on the proximal policy optimization (PPO) algorithm is proposed to jointly optimize transmission power and power allocation coefficients, maximizing achievable data rates under power constraints. The DRL-based approach offers adaptability, low training overhead, and efficiency in dynamic environments.",

keywords = "Reconfigurable intelligent surface (RIS), Short packet communications (SPC), Deep reinforcement learning (DRL)",

author = "Waqas Khalid and Heejung Yu",

year = "2025",

month = apr,

day = "23",

language = "English",

booktitle = "35th Joint Conference on Communications and Information, Sokcho, South Korea",

}

TY - GEN

T1 - RIS-Aided NOMA Downlink Networks: A DRL-Based Approach

AU - Khalid, Waqas

AU - Yu, Heejung

PY - 2025/4/23

Y1 - 2025/4/23

N2 - This paper addresses the power allocation problem in a reconfigurable intelligent surface (RIS)- aided non-orthogonal multiple access (NOMA) downlink system for short packet communications (SPC). A deep reinforcement learning (DRL) framework based on the proximal policy optimization (PPO) algorithm is proposed to jointly optimize transmission power and power allocation coefficients, maximizing achievable data rates under power constraints. The DRL-based approach offers adaptability, low training overhead, and efficiency in dynamic environments.

AB - This paper addresses the power allocation problem in a reconfigurable intelligent surface (RIS)- aided non-orthogonal multiple access (NOMA) downlink system for short packet communications (SPC). A deep reinforcement learning (DRL) framework based on the proximal policy optimization (PPO) algorithm is proposed to jointly optimize transmission power and power allocation coefficients, maximizing achievable data rates under power constraints. The DRL-based approach offers adaptability, low training overhead, and efficiency in dynamic environments.

KW - Reconfigurable intelligent surface (RIS)

KW - Short packet communications (SPC)

KW - Deep reinforcement learning (DRL)

UR - https://jcci.kr/wp-content/uploads/2025/04/JCCI2025-%ED%94%84%EB%A1%9C%EA%B7%B8%EB%9E%A8%EB%B6%81_v17.pdf

M3 - Conference contribution

BT - 35th Joint Conference on Communications and Information, Sokcho, South Korea

ER -

RIS-Aided NOMA Downlink Networks: A DRL-Based Approach

Abstract

Keywords

Other files and links

Fingerprint

Cite this