Research Of Spectrum Allocation For Multi-cells In C-RAN

There is a typical tidal-effect in mobile communication networks, namely that network traffic varies obviously at different time of a day. The utilization of network resource is low because the current networks are mainly designed and built to meet the maximal demand of users. All spectrum resources are managed by a base band processing unit in a centralized manner so as to improve spectrum efficiency in the C-RAN network.This thesis studies spectrum resource allocation in C-RAN network. The main works are as follows:1. Based on the Markov transition model, this thesis designs a prediction scheme of cell users named HCPTI. Based on the simplification of the mobile model and queuing theory, a Markov transition model of the number of cell's users is built. The arrival rate and service rate are analyzed and the values are obtained. In order to obtain the expected value of the number of users in every cell, the transient solution of the queuing models at different time is solved. Considering prediction error and system overhead, the optimal forecast period is obtained. The expected value of the number of users in one cell is obtained by using the optimal forecast period. Comparing with AR, the simulation results show that the scheme presented in this thesis outperforms AR.2. Based on HCPTI, this thesis designs an initial spectrum resource allocation scheme and the corresponding adjustment scheme, namely FABDR and INRA, respectively. The bandwidth requirement is obtained at different time according to the probability of daily spectrum utilization. Considering the size of carrier, the utilization of spectrum and the number of allocation times, the minimum allocation unit is defined. According to spectrum requirements of cells, initial allocation is conducted at the beginning of every period by FABDR. After the initial allocation, the adjustment is conducted by INRA if the spectrum supply can't meet the corresponding demands. Considering the utilization of spectrum, throughout, blocking rate and energy efficiency, the simulation results show that the presented scheme is better than fixed allocation scheme and periodical batch allocation scheme.