Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Joint QoS-Aware Scheduling and Precoding for Massive MIMO Systems via Deep Reinforcement Learning

Chih-Wei Huang, Yen-Cheng Chou, Hong-Yunn Chen, Cheng-Fu Chou

TL;DR

This work tackles QoS-aware joint resource management in massive MIMO by formulating the problem as a Markov decision process and solving it with deep reinforcement learning. It introduces a componentized action structure and action embedding to enable dynamic algorithm selection across user prioritization, antenna allocation, and hybrid precoding, guided by a QoS-integrated long-term utility. The proposed DDGP-based framework demonstrates improvements in the number of satisfied users (e.g., 7.2% and 12.5% over static and baselines) under demanding traffic mixes, validating its effectiveness and adaptability. The approach offers a scalable, cross-layer solution for next-generation networks where QoS requirements and massive resource flexibility coexist.

Abstract

The rapid development of mobile networks proliferates the demands of high data rate, low latency, and high-reliability applications for the fifth-generation (5G) and beyond (B5G) mobile networks. Concurrently, the massive multiple-input-multiple-output (MIMO) technology is essential to realize the vision and requires coordination with resource management functions for high user experiences. Though conventional cross-layer adaptation algorithms have been developed to schedule and allocate network resources, the complexity of resulting rules is high with diverse quality of service (QoS) requirements and B5G features. In this work, we consider a joint user scheduling, antenna allocation, and precoding problem in a massive MIMO system. Instead of directly assigning resources, such as the number of antennas, the allocation process is transformed into a deep reinforcement learning (DRL) based dynamic algorithm selection problem for efficient Markov decision process (MDP) modeling and policy training. Specifically, the proposed utility function integrates QoS requirements and constraints toward a long-term system-wide objective that matches the MDP return. The componentized action structure with action embedding further incorporates the resource management process into the model. Simulations show 7.2% and 12.5% more satisfied users against static algorithm selection and related works under demanding scenarios.

Joint QoS-Aware Scheduling and Precoding for Massive MIMO Systems via Deep Reinforcement Learning

TL;DR

This work tackles QoS-aware joint resource management in massive MIMO by formulating the problem as a Markov decision process and solving it with deep reinforcement learning. It introduces a componentized action structure and action embedding to enable dynamic algorithm selection across user prioritization, antenna allocation, and hybrid precoding, guided by a QoS-integrated long-term utility. The proposed DDGP-based framework demonstrates improvements in the number of satisfied users (e.g., 7.2% and 12.5% over static and baselines) under demanding traffic mixes, validating its effectiveness and adaptability. The approach offers a scalable, cross-layer solution for next-generation networks where QoS requirements and massive resource flexibility coexist.

Abstract

The rapid development of mobile networks proliferates the demands of high data rate, low latency, and high-reliability applications for the fifth-generation (5G) and beyond (B5G) mobile networks. Concurrently, the massive multiple-input-multiple-output (MIMO) technology is essential to realize the vision and requires coordination with resource management functions for high user experiences. Though conventional cross-layer adaptation algorithms have been developed to schedule and allocate network resources, the complexity of resulting rules is high with diverse quality of service (QoS) requirements and B5G features. In this work, we consider a joint user scheduling, antenna allocation, and precoding problem in a massive MIMO system. Instead of directly assigning resources, such as the number of antennas, the allocation process is transformed into a deep reinforcement learning (DRL) based dynamic algorithm selection problem for efficient Markov decision process (MDP) modeling and policy training. Specifically, the proposed utility function integrates QoS requirements and constraints toward a long-term system-wide objective that matches the MDP return. The componentized action structure with action embedding further incorporates the resource management process into the model. Simulations show 7.2% and 12.5% more satisfied users against static algorithm selection and related works under demanding scenarios.

Paper Structure

This paper contains 19 sections, 29 equations, 7 figures, 5 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related Works
  3. Joint User Scheduling and Precoding
  4. Resource Management with Machine Learning
  5. Deep Reinforcement Learning and DDPG
  6. System Model and Problem Formulation
  7. System Model
  8. Hybrid Precoding
  9. User Scheduling and Antenna Allocation
  10. QoS-Aware Joint Resource Allocation
  11. Deep Reinforcement Learning for Joint Massive MIMO Resource Allocation
  12. Markov Decision Process Formulation
  13. Componentized Actions
  14. Action Embedding and Training Procedures
  15. Numerical Results
  16. ...and 4 more sections

Figures (7)

  • Figure 1: Joint scheduling and precoding massive MIMO model in three function components: user prioritization, antenna allocation, and precoding.
  • Figure 2: DDPG structure for massive MIMO resource allocation.
  • Figure 3: Componentized actions and fundamental methods in each component.
  • Figure 4: Compare with static methods.
  • Figure 5: Action choose rate in DDPG agent
  • ...and 2 more figures