長程演進技術(Long Term Evolution, LTE)是由電信相關業者所組成的3GPP組織所制訂之無線通訊傳輸標準之一,LTE在下行鏈路中採用正交分頻多重存取(Orthogonal Frequency-Division Multiple Access,OFDMA)的調變技術,相較於過去電信技術如Universal Mobile Telecommunications System (UMTS)所使用的分碼多重存取(Code Division Multiple Access,CDMA),OFDMA可以提供更大的吞吐量,配合2x2 4x4 Multi-Input Multi-Output (MIMO)的天線技術,並可同時支援TDD與FDD模式,可以大幅提高無線技術的傳輸效能。 LTE技術現在日趨成熟,也越來越有未來4G網路標準的趨勢,當今有許多研究主張將Proportionally Fair (PF)以及Max Rate帶入LTE中,基於PF演算法的公平性以及Max Rate演算法最大化吞吐量的特性,但是當Max Rate演算法實際運用在LTE的Guaranteed Bit Rate(GBR) traffic上時,因為有QoSClass Identifier (QCI)的優先權影響,發現其可能在特定情況下無法達到最大化吞吐量的目的,因此我們將傳統的Max Rate演算法做改變,並且將其命名為EnhancedUtilization Resource Allocation Scheme in GBR traffic(EURA_GBR),使其運用在LTE上時,依然能達到最大化吞吐量的目的,最後我們將傳統的PF演算法運作在Non-Guaranteed Bit Rate (non-GBR) traffic上時只針對不同User Equipment (UE) 間的Rate的公平性這項特性去做延伸,提出一套可以達到在相同UE底下不同flow間根據Packet Delay Budget(PDB)的延遲公平性演算法,稱之為Delay-based PF.
Long Term Evolution (LTE) is one of 3GPP standards developed by 3GPP which is organized by telecommunication manufactories and operators. LTE uses Orthogonal Frequency Division Multiple Access (OFDMA) technology on the downlink. The OFDMA can have higher throughput than Code Division Multiple Access (CDMA) in Universal Mobile Telecommunications System (UMTS). Based on 2x2 and 4x4 Multi-Input Multi-Output (MIMO) antenna system, LTE can support both TDD and FDD modes to achieve higher peak data rates. Many researches tried to adopt Max Rate or Proportionally Fair (PF) algorithm as LTE MAC scheduling scheme in eNB due to maximization of throughput or bandwidth fairness individually. Considering LTE MAC QoS requirement and the variation Channel Quality Indicator (CQI) in UEs, the original Max Rate algorithm can not achieve the goal of maximization throughput. In the thesis, an Enhanced Utilization Resource Allocation Scheme in GBR traffic (EURA_GBR) with three methods is proposed. The proposed scheme can improve the utilization of LTE radio resource. Moreover, a Delay based PF algorithm in non-GBR traffic is also proposed to achieve delay fairness target.