认知无线电网络中最优分组长度设计

为满足不同种类多媒体应用的服务质量需求,研究了一个多等级多信道认知无线电网络中的资源分配问题。根据时延和吞吐量为该网络模型定义一个效用函数并最大化这个效用函数的期望值。首先用不考虑传输功率分配的"浪费功率的方式"设计分组长度。然后在选择分组长度时考虑传输功率分配提出一种基于贪婪算法的改进方式,得到了最优分组长度和相应的传输功率。MATLAB的仿真结果表明,在传输功率受限下,当达到相同效用值时,改进方式和"浪费功率的方式"相比能减小传输功率,并取得更加合理的分组长度,而且能更好地满足带宽需求,这是由于改进方式不仅考虑了各个信道上数据包的到达时间间隔,还考虑了信道增益。

Optimal Packet Length Design in Cognitive Radio Network

To satisfy the QoS(quality of service)requirements of different kinds of multimedia applications,we study problems of resource allocation in a multi-class multi-channel cognitive radio network.We define a utility function for the network model according to users' delay and throughput and aims at maximizing the expected utility function.We firstly design packet lengths using the"power-consuming"method without considering transmission power allocation.Then we propose an improved method based on the greedy algorithm that considers transmission power allocation during the packet lengths selection to obtain the optimal packet lengths and the corresponding transmission power.The simulation results in MATLAB demonstrate that under the limited transmission power,the improved method can decrease the transmitting power when achieving the same utility compared to"power-consuming"method,and it can obtain more reasonable packet lengths and better meet the bandwidth demand because it considers not only intervals between packets' arrival on each channel but also channel gains.