Abstract
The rapid evolution of mobile telecommunication networks has intensified the demand
for enhanced energy efficiency, scalability, coverage, and service quality. Despite
ongoing advances in automation and artificial intelligence, existing network management
systems continue to face challenges in maintaining the optimal balance between
performance, energy consumption, and operational reliability—particularly in dynamic,
data-intensive environments such as fifth-generation (5G) and beyond networks. These
limitations highlight the need for an adaptive, interpretable, and self-governing control
framework capable of managing network functions autonomously while ensuring
transparency and robustness.
This research characterises and develops the Normative Reinforcement Control
Framework (NRCF)—a hybrid theoretical and practical framework integrating normative
control principles with reinforcement learning to enhance autonomy, decision integrity,
and explainability in autonomous network functions (ANFs). The study systematically
evaluates the performance and maturity of ANFs through a novel medically inspired
measurement system (GANS, MEGABITS, and Gleason-style scoring) and introduces
advanced anomaly detection models, ANOBIA and INFEROBIA, for reliable bias
identification and correction. Empirical evaluations demonstrate that the NRCF achieves
consistent network optimisation with a prediction accuracy of 90%, improving energy
efficiency while preserving service reliability under variable conditions.
Furthermore, this research proposes the Intelligent Plane, a new architectural layer for
beyond-5G networks, designed to centralise AI-driven governance, ensure coordinated
decision-making, and embed explainability and ethical reasoning directly into network
control systems. The findings establish the NRCF and Intelligent Plane as foundational
elements for intelligent, sustainable, and transparent network architectures.
The rapid evolution of mobile telecommunication networks has intensified the demand
for enhanced energy efficiency, scalability, coverage, and service quality. Despite
ongoing advances in automation and artificial intelligence, existing network management
systems continue to face challenges in maintaining the optimal balance between
performance, energy consumption, and operational reliability—particularly in dynamic,
data-intensive environments such as fifth-generation (5G) and beyond networks. These
limitations highlight the need for an adaptive, interpretable, and self-governing control
framework capable of managing network functions autonomously while ensuring
transparency and robustness.
This research characterises and develops the Normative Reinforcement Control
Framework (NRCF)—a hybrid theoretical and practical framework integrating normative
control principles with reinforcement learning to enhance autonomy, decision integrity,
and explainability in autonomous network functions (ANFs). The study systematically
evaluates the performance and maturity of ANFs through a novel medically inspired
measurement system (GANS, MEGABITS, and Gleason-style scoring) and introduces
advanced anomaly detection models, ANOBIA and INFEROBIA, for reliable bias
identification and correction. Empirical evaluations demonstrate that the NRCF achieves
consistent network optimisation with a prediction accuracy of 90%, improving energy
efficiency while preserving service reliability under variable conditions.
Furthermore, this research proposes the Intelligent Plane, a new architectural layer for
beyond-5G networks, designed to centralise AI-driven governance, ensure coordinated
decision-making, and embed explainability and ethical reasoning directly into network
control systems. The findings establish the NRCF and Intelligent Plane as foundational
elements for intelligent, sustainable, and transparent network architectures.