5 G毫米波網路中以比例公平為基礎之用戶連線與資源配置

Abstract

5G毫米波網絡使用在頻段為GHz，已知能夠提供Gbps級別的速度。然而，在這些頻段中，通道容易受到損壞。因此，類似Wi-Fi網絡，小型基站被部署在服務區域內，以解決這個問題。本論文討論了一種基於對偶分解方法的分散式演算法。在該分散式演算法中，不僅可以確定用戶端和存取點之間的連線，還可以以一種比例公平(proportional fairness,PF)的方式分配通道資源。具體而言，本論文運用隨機幾何學來計算網絡整體吞吐性能的平均值。這有助於節省從大量模擬運行中取平均值的工作。

A 5G millimeter network, operated in the bands of GHz, is known to provide speeds of Gbps. However, channel is prone to be deteriorated in these bands. Therefore, like a Wi-Fi network, small cells are deployed in a service area to combat this problem.In this thesis, a distributed algorithm based on the dual decomposition method is discussed. In the distributed algorithm, not only can the associations between clients and access points (APs) be determined, but channel resources are allocated in a way that proportional fairness (PF) can be achieved. In particular, stochastic geometry is employed in the thesis to calculate the mean of the overall throughput performance of the network. This helps to spare the work of taking the average from a lot of simulation runs.