Neural Network-Based Adaptive Resource Allocation for 5G Heterogeneous Ultra-Dense Networks

Area of Science:

• Telecommunications Engineering
• Wireless Network Optimization
• Artificial Intelligence in Networks

Background:

• 5G networks face challenges with increasing traffic demands despite wider spectral bandwidth.
• Heterogeneous ultra-dense networks (HUDNs) are crucial for traffic offloading but present complex resource allocation challenges due to diverse base stations and Quality of Service (QoS) needs.
• Static resource allocation methods are inflexible and lead to inefficient spectrum use in complex 5G environments.

Purpose of the Study:

• To develop a joint user association-resource allocation (UA-RA) framework for 5G HUDNs.
• To dynamically adapt resource allocation to real-time network conditions for improved spectral efficiency and service ratio under high traffic loads.
• To mitigate congestion through inter-cell resource redistribution managed by a software-defined networking (SDN) controller.

Main Methods:

• A joint user association-resource allocation (UA-RA) framework was developed for 5G HUDNs.
• A software-defined networking (SDN) controller was utilized for centralized UA-RA and inter-cell resource redistribution.
• A neural network-adaptive resource allocation (NN-ARA) model was trained on simulation data to approximate efficient allocation decisions with low computational cost.

Main Results:

• The NN-ARA approach demonstrated significant improvements compared to baseline methods.
• Achieved up to 20.8% and 11% higher downlink data rates in macro and small cell tiers, respectively.
• Improved spectral efficiency by approximately 20.7% and 11.1%, and reduced the average blocking ratio by up to 55%.

Conclusions:

• The NN-ARA framework offers an adaptive, scalable, and SDN-coordinated solution for efficient spectrum utilization in 5G and future 6G HUDNs.
• The proposed model effectively addresses the challenges of dynamic resource allocation in complex network environments.
• NN-ARA ensures improved service continuity and network performance under high traffic loads.