Joint beam allocation method based on Bayesian parameter adjustment support vector machine

Abstract

The invention provides a beam allocation method of a support vector machine based on Bayesian parameter modulation, which comprises the following steps: firstly, generating a to-be-selected beam allocation scheme space, and then randomly generating an initial beam allocation scheme set or generating an initial beam allocation scheme set according to selection schemes in similar historical allocation records; the method being applied to a cellular mobile communication system to obtain evaluation of the system on a scheme, and labeling being performed to generate training samples; performing parameter optimization by using a Bayesian hyper-parameter optimization method in combination with cross validation, and then learning a training sample in a (scheme, label) format by using a support vector machine machine learning algorithm to obtain a corresponding constraint condition; and finally, screening a beam allocation scheme in a beam scheme space by using a beam scheme optimization algorithm and combining the obtained constraint conditions to generate a next beam allocation scheme set. According to the invention, the technical problems of low beam allocation matching degree and low allocation efficiency in the prior art are solved, and high real-time performance is achieved while the accuracy is ensured.

Description

technical field [0001] The present invention relates to the technical field of wireless communication, specifically proposes a multi-cell interference sensing and joint beam allocation method, combines support vector machines and mobile communications, and provides a beam allocation method based on support vector machines in daily traffic scenarios . Background technique [0002] In recent years, with the deepening of machine learning research, machine learning algorithms have been widely used to solve problems in various fields, such as speech recognition, image recognition, etc., which shows that machine learning algorithms have strong universality. In the field of wireless communication research, machine learning algorithms can also be used to solve traditional communication problems. Both beamforming and signal detection have many successful research applications. [0003] Compared with other machine learning algorithms, support vector machine (SVM) can usually achieve ...

AI Technical Summary

Problems solved by technology

[0004] In order to solve the problem of low beam allocation efficiency and high computational complexity of traditional optimization methods, the present invention proposes a joint beam allocation method based on Bayesian parameter adjustment support vector machine

Images