期刊
INTERNATIONAL JOURNAL OF ELECTRONICS
卷 108, 期 7, 页码 1214-1233出版社
TAYLOR & FRANCIS LTD
DOI: 10.1080/00207217.2020.1843715
关键词
5G mMIMO; multi-objective functions; MOSCA; energy efficiency; DNN
5G network is crucial in various fields and user numbers are increasing. Machine learning methods are introduced for better resource allocation in 5G networks. The proposed method using deep neural network for resource allocation in 5G mMIMO outperforms existing methods.
5G network plays a vital role in each field. Recently, the number of 5G users are increasing due to their vast merits. But, these users require various resources to operate effectively. Recently, the machine learning approaches are introduced within this 5G network field for attaining better accuracy and reliability during resource allocation. In this work, the resource allocation for multi-users in 5G massive-MIMO (mMIMO) is performed by a deep neural network (DNN). Initially, the objective functions are optimised by Multi-objective Sine Cosine algorithm (MOSCA). The objective functions applied in this optimisation method are data rate, signal-interference noise ratio (SINR), power consumption, and energy efficiency (EE). Next, these optimised objective functions are allocated to the neural network for resource allocation. DNN identifies the requirement level of each user. Based on this level, it allocates the resource to each user by maintaining high throughput and EE. Further, the fairness index for this neural network-based resource allocation process is also identified. The throughput and fairness index of the proposed method for 30 users are 275 bps, and 0.92%, respectively. The outcomes of this proposed method show that this method provides better performance in 5G mMIMO than other existing methods.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据