Proactive buffer control method in wireless communication network based on learning

Abstract

The invention discloses a proactive buffer control method in a wireless communication network based on learning, and belongs to the technical field of wireless communication. A proactive buffer technology is adopted by a proactive buffer scheme in the wireless communication network based on learning, and by buffering network edge equipment, predictable user requirements can be proactively met. When the popularity degree between a file and a user is predicted and estimated, aiming at the problem that in reality, a file popularity matrix is generally very large and sparse, collaborative filtering and transfer learning are utilized to obtain a more accurate estimation result. When a buffer strategy of a small base station is formulated, a distributed iteration algorithm is designed to betterexplore a globally optimal solution of the buffer. By means of the proactive buffer scheme, the user data rate can be effectively increased, and the return load can be effectively lowered.

Description

technical field [0001] The invention relates to the technical field of wireless communication, in particular to a learning-based active buffer control method in a wireless communication network. Background technique [0002] With the development of mobile communications, mobile smart terminals and social networks, we are increasingly using mobile terminals to experience a variety of online services. The improvement of the level of user demand has driven the development of a large number of emerging wireless communication services, and also brought about the explosive growth of the data volume of the mobile Internet. To cope with the explosive growth of mobile data volumes today, our traditional approach is to increase spectral resources (bandwidth), spectral efficiency (modulation, coding, MIMO), or spatial multiplexing (base station density). However, under practical application conditions, these methods provide rather limited throughput gain, [0003] And the implementat...

AI Technical Summary

Problems solved by technology

However, in reality, the file popularity matrix is ​​usually very large and sparse, which makes CF technology perform poorly on issues such as data sparsity and cold start, prompting us to seek better learning methods to improve the performance of the caching scheme

In addition, for the caching strategy of small base stations, greedy caching of the most popular files results in a local optimum. Currently, there is no better caching strategy for small base stations to obtain the global optimal solution.

Images