An Adaptive Learning Based Network Selection Approach for 5G Dynamic Environments.

Area of Science:

• Computer Science
• Telecommunications Engineering

Background:

• Future 5G networks will be highly heterogeneous with dynamically changing radio resources.
• Existing network selection methods are inadequate for these dynamic and unpredictable 5G environments.

Purpose of the Study:

• To develop an adaptive learning strategy for terminal users to select optimal access networks.
• To address the challenges posed by fluctuating radio resources and varying user demands in heterogeneous networks.

Main Methods:

• Modeling the network selection as a multiagent coordination problem with rational users.
• Proposing an adaptive learning strategy enabling users to adjust selections dynamically.
• Analyzing system convergence to Nash equilibrium, Pareto optimality, and social optimality.

Main Results:

• The proposed adaptive learning strategy significantly outperforms existing methods.
• Demonstrated improvements in load balancing, user payoff, and bandwidth utilization efficiency.
• The system exhibits robust performance even with non-compliant users.

Conclusions:

• The adaptive learning strategy effectively manages dynamic network conditions and user demands in 5G.
• The approach ensures optimal network selection, leading to enhanced system performance and user satisfaction.
• The strategy offers a robust solution for future heterogeneous network environments.