LTE-A 異質網路下協同排程與資源分配之聯合最佳化

Abstract

在下一代通訊網路下，資源分配及排程是兩個關鍵性的議題。不像過去的相 關研究多半注重在同質性無線通訊系統下進行資源分配及排程，在本論文裡，我 們考慮一個包含傳統大型基地台以及數個低功率遠端射頻收發裝置的異質性無 線通訊系統，並研究如何經由各異質傳輸點之間的協同排程及資源分配，以提供 所有使用者公平而有效率的服務。為了設計出最佳之資源分配以及排程，我們首 先將問題抽象化成一個最佳化問題，藉由綜合考慮各個排程時間及傳輸點頻譜資 源以達到一個對所有使用者公平而最佳的分配。然而，由於這個方法的搜尋複雜 度太高，因此我們接著將問題近似成一個疊代式的排程問題，此方法將原先的聯 合最佳化問題拆解成若干個子問題，並將目標函數經過適當轉化以降低數學運算 的複雜度。疊代式的排程雖然在在最初的排程效能有一些損失，但隨著排程時間 的推展，其解法可極度趨近於原先的問題，達到公平而有效率的資源分配。基於 最佳化的資源分配及協同排程求解，我們最後提出一個低複雜度的演算法，藉由 觀察各個使用者的效能高低以及通道環境的變動，動態進行資源分配，確保在達 到公平的前提下能伴隨著最少的通道容量損失。電腦模擬結果顯示，不管就使用 者公平性或通道效能而言，此演算法都明顯優於相關文獻的作法；與前述最佳化 問題之最佳解比較，此演算法也僅需極短的時間即可趨近於最佳解，達到異質性 無線通訊系統下協同排程與資源分配之目的。

Resource allocation and transmission scheduling are two key issues in the next-generation communication system. Unlike related work that focuses mostly on homogeneous networks, we consider in this thesis a heterogeneous network with macro BSs and several low-power remote radio heads (RRHs). The goal is to perform coordinated scheduling among heterogeneous transmission points through proper resource allocation for all users in the OFDMA-based system. To proceed, we first formulate an optimization problem that takes into consideration resource allocation and transmission scheduling across all scheduling slots. Since the computation complexity is prohibitive, we then transform the one-shot optimization problem into an iterative scheduling problem. While the transformed problem exhibits some performance degradation initially, as the amount of scheduling slots increases, its performance improves and closely tracks the original formulation with significantly lower complexity. Based on insights from solving the optimization problem, we then propose a coordinated scheduling algorithm to dynamically allocate resources of all transmission points to users based on the perceived performance and channel variation. Through proper control of fairness and aggregate capacity, the algorithm can achieved the desired performance and outperforms approaches proposed in related work. We thus motivate further investigation for joint coordinated scheduling and resource allocation across heterogeneous transmission points in the next-generation wireless system.