Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Toward hyper-adaptive AI-enabled 6G networks for energy efficiency: techniques, classifications and tradeoffs

Mariem Zayene, Oussama Habachi, Gerard Chalhoub

TL;DR

This work tackles the challenge of achieving energy efficiency in 6G networks through AI-driven adaptability. It introduces six EE-first use cases and a framework to evaluate AI methods under seven core 6G dynamics, emphasizing adaptability and tradeoffs across mobility, channel variability, traffic, topology, and service heterogeneity. The survey consolidates AI techniques across RIS, UAV, IIoT/smart city, V2X, and edge intelligence, highlighting that reinforcement learning and multi-agent systems provide the strongest real-time adaptability while digital twins and prediction enable proactive planning; yet hybrid, cross-layer approaches are necessary to balance energy savings with QoS, fairness, and coverage. The findings underscore the need for generalizable, low-overhead, and explainable AI that can operate across domains, scales, and dynamic conditions, making energy-aware AI a foundational pillar of sustainable 6G networks.

Abstract

Energy efficiency is shaping up to be one of the most challenging issues for 6G networks. The reason is fairly straightforward: Networks will need to meet extreme service demands while remaining sustainable and traditional optimization techniques are too limited. With users moving, traffic swinging unpredictably and services pulling in different directions, management has to be adaptive and AI may offer a way forward. This survey looks at how well AI-based methods actually deliver on that promise. We organize the review around practical use cases. For each use case, we examine how AI techniques contribute to feedback-driven adaptability and rapid decision-making under dynamic conditions. We then evaluate them against seven central dynamic aspects that we consider unavoidable in 6G. The survey also discusses crucial tradeoffs between energy efficiency and the remaining 6G main objectives such as latency, reliability, fairness and coverage, and finally identifies gaps and future research directions.

Toward hyper-adaptive AI-enabled 6G networks for energy efficiency: techniques, classifications and tradeoffs

TL;DR

This work tackles the challenge of achieving energy efficiency in 6G networks through AI-driven adaptability. It introduces six EE-first use cases and a framework to evaluate AI methods under seven core 6G dynamics, emphasizing adaptability and tradeoffs across mobility, channel variability, traffic, topology, and service heterogeneity. The survey consolidates AI techniques across RIS, UAV, IIoT/smart city, V2X, and edge intelligence, highlighting that reinforcement learning and multi-agent systems provide the strongest real-time adaptability while digital twins and prediction enable proactive planning; yet hybrid, cross-layer approaches are necessary to balance energy savings with QoS, fairness, and coverage. The findings underscore the need for generalizable, low-overhead, and explainable AI that can operate across domains, scales, and dynamic conditions, making energy-aware AI a foundational pillar of sustainable 6G networks.

Abstract

Energy efficiency is shaping up to be one of the most challenging issues for 6G networks. The reason is fairly straightforward: Networks will need to meet extreme service demands while remaining sustainable and traditional optimization techniques are too limited. With users moving, traffic swinging unpredictably and services pulling in different directions, management has to be adaptive and AI may offer a way forward. This survey looks at how well AI-based methods actually deliver on that promise. We organize the review around practical use cases. For each use case, we examine how AI techniques contribute to feedback-driven adaptability and rapid decision-making under dynamic conditions. We then evaluate them against seven central dynamic aspects that we consider unavoidable in 6G. The survey also discusses crucial tradeoffs between energy efficiency and the remaining 6G main objectives such as latency, reliability, fairness and coverage, and finally identifies gaps and future research directions.

Paper Structure

This paper contains 47 sections, 2 figures, 7 tables.

Table of Contents

  1. Introduction
  2. From 1G to 6G: A progressive path toward network adaptability
  3. Visionary architectures for adaptive 6G networks
  4. Nokia – Hexa-X 6G Vision.
  5. Ericsson – 6G RAN Energy Vision.
  6. Huawei – "6G: The Next Horizon" Vision (2023).
  7. Why an adaptability focused AI survey is needed?
  8. Related Work
  9. Dedicated AI-for-EE Surveys
  10. Domain-specific surveys (partial EE focus)
  11. Vision and framework-oriented surveys
  12. General 6G surveys with broad scope
  13. Positioning of this survey and contributions
  14. Energy efficiency and sustainability in 6G use cases: From standardized visions to EE-first lenses
  15. Why is energy efficiency crucial for 6G network?
  16. ...and 32 more sections

Figures (2)

  • Figure 1: The 6G EE-first lenses
  • Figure 2: The inherent 6G dynamic aspects affecting EE