Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Feasibility-Aware Learning-to-Optimize in Wireless Communication Resource Allocation

Hanwen Zhang, Haijian Sun

TL;DR

This work examines feasibility-aware learning-to-optimize (L2O) for real-time wireless resource allocation (RA) in 6G, where high-dimensional, constrained decision problems arise in IoT and V2X contexts. It synthesizes data-driven and model-driven L2O designs, including deep unfolding (DU), graph neural networks (GNN), and neural building blocks, and categorizes feasibility enforcement into soft (penalty-based, algorithmic) and hard (iterative, structural) approaches. A QoS-aware $WSR$ case study on a downlink MU-MISO system demonstrates that hard-constrained, model-driven architectures—especially a DC3+DU configuration—achieve superior performance with near-zero constraint violations, highlighting the importance of integrating optimization structure with feasibility layers. The paper concludes with strategic directions to expand RA applications, tailor task-specific L2O models, address multi-variable dependencies, and embed constraints into neural architectures, aiming to enable scalable, differentiable, and reliable real-time RA for future wireless systems.

Abstract

The emergence of 6G wireless communication enables massive edge device access and supports real-time intelligent services such as the Internet of things (IoT) and vehicle-to-everything (V2X). However, the surge in edge devices connectivity renders wireless resource allocation (RA) tasks as large-scale constrained optimization problems, whereas the stringent real-time requirement poses significant computational challenge for traditional algorithms. To address the challenge, feasibility-aware learning-to-optimize (L2O) techniques have recently gained attention. These learning-based methods offer efficient alternatives to conventional solvers by directly learning mappings from system parameters to feasible and near-optimal solutions. This article provide a comprehensive review of L2O model designs and feasibility enforcement techniques and investigates the application of constrained L2O in wireless RA systems and. The paper also presents a case study to benchmark different L2O approaches in weighted sum rate problem, and concludes by identifying key challenges and future research directions.

Feasibility-Aware Learning-to-Optimize in Wireless Communication Resource Allocation

TL;DR

This work examines feasibility-aware learning-to-optimize (L2O) for real-time wireless resource allocation (RA) in 6G, where high-dimensional, constrained decision problems arise in IoT and V2X contexts. It synthesizes data-driven and model-driven L2O designs, including deep unfolding (DU), graph neural networks (GNN), and neural building blocks, and categorizes feasibility enforcement into soft (penalty-based, algorithmic) and hard (iterative, structural) approaches. A QoS-aware case study on a downlink MU-MISO system demonstrates that hard-constrained, model-driven architectures—especially a DC3+DU configuration—achieve superior performance with near-zero constraint violations, highlighting the importance of integrating optimization structure with feasibility layers. The paper concludes with strategic directions to expand RA applications, tailor task-specific L2O models, address multi-variable dependencies, and embed constraints into neural architectures, aiming to enable scalable, differentiable, and reliable real-time RA for future wireless systems.

Abstract

The emergence of 6G wireless communication enables massive edge device access and supports real-time intelligent services such as the Internet of things (IoT) and vehicle-to-everything (V2X). However, the surge in edge devices connectivity renders wireless resource allocation (RA) tasks as large-scale constrained optimization problems, whereas the stringent real-time requirement poses significant computational challenge for traditional algorithms. To address the challenge, feasibility-aware learning-to-optimize (L2O) techniques have recently gained attention. These learning-based methods offer efficient alternatives to conventional solvers by directly learning mappings from system parameters to feasible and near-optimal solutions. This article provide a comprehensive review of L2O model designs and feasibility enforcement techniques and investigates the application of constrained L2O in wireless RA systems and. The paper also presents a case study to benchmark different L2O approaches in weighted sum rate problem, and concludes by identifying key challenges and future research directions.

Paper Structure

This paper contains 27 sections, 5 figures, 1 table.

Table of Contents

  1. Introduction
  2. Fundamentals of Learning to Optimize in Constrained Optimization
  3. Overview of L2O Mechanism
  4. NN Designing for Different Application Requirement
  5. Multilayer Perceptron (MLP)
  6. Graph Neural Network (GNN)
  7. Deep Unfolding (DU)
  8. Neural Building Blocks
  9. Specialization for L2O Training Approaches
  10. Supervised Learning (SL)
  11. Self-supervised Learning (SSL) / Unsupervised Learning (UL)
  12. feasibility-Aware L2O for Constrained Optimization
  13. Soft-constrained L2O
  14. Penalty-based
  15. Algorithmic
  16. ...and 12 more sections

Figures (5)

  • Figure 1: The Overview of L2O Structure
  • Figure 2: Two Perspectives of Feasibility Aware L2O
  • Figure 3: Feasibility-Aware L2O Methods
  • Figure 4: Constrained L2O Performance over Different Channel SNRs
  • Figure 5: Constrained L2O Performance over Different Nagakami-m Channel