Distributed Resources Allocation Method for Space-Ground Integrated Mobile Communication System.

Area of Science:

• Communications Engineering
• Network Architecture
• Artificial Intelligence in Networks

Background:

• Future communication systems require seamless integration of terrestrial, UAV, and satellite networks.
• Increasing user demands and network complexity necessitate innovative resource management strategies.
• Existing network architectures face challenges in handling dynamic user surges and reducing overhead.

Purpose of the Study:

• To propose an innovative space-ground integrated communication system architecture.
• To develop an efficient network resource allocation decision-making model for mobile stations (MSs).
• To reduce core network overhead and improve the success access ratio in dynamic environments.

Main Methods:

• Leveraging advanced network slicing technology for integrated networks.
• Introducing a marginal mobile station (MS)-assisted network resource allocation decision architecture.
• Developing and applying a Deep Policy Q-Network (DP-DQN) model for enhanced MS decision-making.
• Implementing a feedback mechanism for continuous model adaptation and accuracy.

Main Results:

• The DP-DQN-based edge decision method significantly alleviates core network overhead.
• Demonstrated improvement in success access ratios compared to conventional methods.
• Validated effectiveness through extensive simulations and experimental testing.

Conclusions:

• The proposed DP-DQN-based edge decision method offers a promising solution for space-ground integrated communication systems.
• The MS-assisted architecture and feedback mechanism enhance network adaptability and efficiency.
• This approach effectively addresses challenges posed by future user surges and network dynamics.