Multi-Agent DRL for User Association and Power Control in Terrestrial-Satellite Network

In the past few years, satellite communications have greatly affected our daily lives. Because the resources of terrestrial-satellite network are limited, how to allocate resources of terrestrial-satellite network through effective methods have become a major challenge. We propose a framework for energy efficiency optimization of terrestrial-satellite network based on Non-orthogonal multiple access (NOMA). In our framework, we adopt a multi-agent deep deterministic policy gradient (MADDPG) method to obtain the maximum energy efficiency by user association and power control. Finally, the simulation results show that the proposed method has better optimization performance compared with the traditional singleagent deep reinforcement learning algorithm and can efficiently solve the problems of user association and power control in the integrated terrestrial-satellite network.

I. Introduction

In recent years, the NOMA schemes are often applied in the integrated terrestrial-satellite network that consists of BSs on the ground and satellites [1]–[3]. It is considered as a promising scenario. As a multiple access technology, NOMA can improve the total energy efficiency of the system [4]–[6]. In the terrestrial-satellite network [7], BSs provide low-cost communication services, while satellites can be used to cover and serve users who are in the underdeveloped areas. This system can achieve wider coverage area and better service quality. With the continuous increase of data traffic in communications, one of the main challenges is how to use effective methods for resource allocation and improve the system energy efficiency.