Deep reinforcement learning-based heterogeneous cellular network joint optimization method

Abstract

The invention belongs to the wireless communication technological field and relates to a deep reinforcement learning-based heterogeneous cellular network joint optimization method. The invention aimsto solve the joint optimization problem of user association, resource allocation and power control in a downlink heterogeneous cellular network. The non-convexity and composite property of the joint optimization problem are considered. A heterogeneous cellular network system model is established. A multi-agent deep reinforcement learning optimization strategy is put forward. The conditions of users' satisfaction and operators' pursuit of interests are considered. A deep reinforcement learning method is used to obtain an optimal resource allocation, user association and power control joint strategy. Thus, the long-term system effectiveness of the entire network can be realized.

Description

technical field [0001] The invention belongs to the technical field of wireless communication, and in particular relates to a heterogeneous cellular network joint optimization method based on deep reinforcement learning. Background technique [0002] With the rapid development of wireless communication technology, heterogeneous cellular networks are facing the huge challenge of increasing network capacity requirements. Heterogeneous cellular networks can balance traffic loads and reduce small device charges, which are considered to be promising technologies in next-generation cellular networks. [0003] At present, there are still many problems in heterogeneous cellular networks, such as user association, resource allocation and power control. Due to the high-dimensional action space of joint optimization problems, it is difficult to obtain the optimal policy using reinforcement learning methods. Considering the high-dimensional action space of joint optimization problems,...

AI Technical Summary

Problems solved by technology

Due to the high-dimensional action space of the joint optimization problem, it is difficult to obtain the optimal policy using reinforcement learning methods

Images