C-RAN网络中多小区频谱资源分配研究

发布时间：2018-04-28 00:34

本文选题：C-RAN + Markov转移模型　；参考：《重庆邮电大学》2016年硕士论文

【摘要】：移动通信网络具有典型的“潮汐效应”,即网络流量在一天中的不同时刻起伏较大。现有通信网络主要按照用户的最大化需求来进行设计和建设,网络资源利用率偏低。在C-RAN网络中,通过基带处理单元对各个小区进行统一管理,为频谱资源的集中化处理提供了条件,以此实现频谱资源的动态分配,提高频谱资源使用效率。本文研究C-RAN网络中的频谱资源分配,主要工作如下:1.设计了一种基于Markov转移模型的小区用户数预测机制,命名为HCPTI。将复杂移动模型进行简化,利用排队论,建立基于小区用户数的Markov转移模型。分析用户到达率和离开率,并对其予以求解。求解排队模型不同时刻的瞬态解,得出不同时刻小区用户数的期望值。综合考虑预测误差和系统开销,求出HCPTI预测机制的最佳预测周期,把最佳预测周期下求得的用户数期望值作为预测结果。在最佳预测周期下把HCPTI预测机制与时间序列自回归预测(AR)方案作进行仿真性能比较,结果表明,HCPTI预测机制的预测性能更优。2.基于小区用户数预测机制,设计了频谱资源初次分配方案(命名为FABDR)和相应的调整分配方案(命名为INRA)。根据用户日常使用网络资源概率的变化情况,求出各个小区在不同时刻的带宽需求。综合考虑载波大小、频谱资源利用率和分配次数这三个因素,确定频谱资源分配的最小分配单元(MAU)。在分配周期的起始时刻对各小区进行频谱资源的初次预分配,当频谱资源供需不匹配时,进行频谱资源分配的调整。综合考察频谱资源利用率、系统容量、系统阻塞率和能耗,把本文分配方案与固定分配方案和周期批处理动态分配方案进行仿真性能比较,结果表明,本文分配方案综合性能更优。
[Abstract]:Mobile communication network has typical tidal effect, that is, the network flow fluctuates greatly at different times of day. The existing communication network is mainly designed and constructed according to the maximum demand of users, and the utilization ratio of network resources is on the low side. In C-RAN network, each cell is managed uniformly by baseband processing unit, which provides the condition for the centralized processing of spectrum resources, so as to realize the dynamic allocation of spectrum resources and improve the efficiency of spectrum resources. In this paper, spectrum resource allocation in C-RAN networks is studied. The main work is as follows: 1. A cell user prediction mechanism based on Markov transfer model is designed and named HCPTI. The complex mobile model is simplified and the Markov transfer model based on the number of cell users is established by using queuing theory. The user arrival rate and departure rate are analyzed and solved. The transient solution of the queuing model at different times is solved, and the expected value of the number of cell users at different times is obtained. Considering the prediction error and system overhead, the optimal prediction period of HCPTI prediction mechanism is obtained, and the expected number of users under the optimal forecasting period is taken as the prediction result. The HCPTI prediction mechanism is compared with the time series autoregressive prediction scheme under the best forecasting period. The results show that the prediction performance of the HCPTI prediction mechanism is better than that of the time series autoregressive prediction scheme. Based on the cell user number prediction mechanism, a spectrum resource allocation scheme (named FABDR) and a corresponding adjusted allocation scheme (named INRAA) are designed. According to the change of the probability of users using network resources, the bandwidth requirements of each cell at different times are calculated. Considering the carrier size, spectrum resource efficiency and allocation times, the minimum allocation unit of spectrum resource allocation is determined. At the beginning of the allocation period, the spectrum resources of each cell are preallocated for the first time, and when the supply and demand of the spectrum resources are not matched, the spectrum resource allocation is adjusted. The spectrum resource efficiency, system capacity, system blocking rate and energy consumption are comprehensively investigated. The simulation performance of the proposed allocation scheme is compared with that of the fixed allocation scheme and the periodic batch dynamic allocation scheme. The results show that, The comprehensive performance of the allocation scheme in this paper is better.
【学位授予单位】：重庆邮电大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TN929.5

【参考文献】