在不同功能切割的雲端無線接取網路架構下之基於服務品質的資源與能量分配機制

Abstract

為了因應通信流量大幅度的成長，利用集中式管理的技術去改善 能量和頻譜的使用效率，儼然成為一個未來通訊網路的趨勢。然而， 集中式的雲端無線電存取網路對於遠端基地台和中央控制器間資料的 傳輸要求非常嚴苛，因此，不得不藉由將無線電存取網路的功能做切 割，將其彈性分配，分別由遠端基地台和中央控制器負責運算處理。 在本篇論文中，考慮三種不同的無線電接入網功能的切割情況，分別 為資源和能量聯合分配，資源和能量分開分配及資源和能量分散分 配，藉由對元件載波，資源區塊和能量的分配方式最佳化系統的能量 效率。但，此能量效率的最佳化問題有著非線性，並且非凸向性的特 性，因此我們藉由隨機性的分配機制演算法去獲得資源分配的策略， 在此同時，亦利用等效的問題轉換，來分析三種情況下系統的表現， 並進一步得出改善第二種無線電存取網路切割的方法。最終，透過模 擬計算，資源和能量分開分配的切割方式不但能減少遠端基地台和中 央控制器間的負荷，並且能藉由論文中所推導出的增強型資源和能量 分開分配的方法來改善其能量效率。

Cloud Radio Access Network (C-RAN) is a promising architecture to enhance the energy efficiency (EE) and spectrum efficiency (SE) through centrally controlling the resource allocation of network system. However, the limited fronthaul capacity makes it implement impractical. Thus, this thesis studies C-RAN with different functional-split architectures to reduce the burden of fronthaul. There are three functional split cases we consider, namely, joint resource and power allocations (J-RPA), separate resource and power allocations (S-RPA), and distributed resource and power allocations (D-RPA). Then formulate the optimization problems to maximize the EE for each case with Component Carrier (CC), Resource Block (RB), and power allocation policies. At the same time, Quality of Service (QoS) requirement for each User Equipment (UE) and limited power budget of Remote Small Cell (RSC) are considered. The problems are non-linear, discrete, and non-convex. The Cross Entropy (CE) algorithm is proposed to address the challenge. In addition, by converting the problems into equivalent ones. And deriving a result to improve the EE performance of S-RPA. For performance evaluation, we compare three functional-split cases under different system requirements, and show that the S-RPA is more powerful than D-RPA and similar performance as J-RPA. Also we propose enhanced S-RPA (ES-RPA) to improve the EE performance of S-RPA.