分布多跳式网络吞吐量优化的并行性研究

On Parallel Studies of Throughput Optimization for Distributed Multi-Hop Networks

传统的分布多跳式网络吞吐量的优化方法并不能满足用户高移动性、高数据速率的要求.为了提高分布多跳式网络吞吐量的优化性能,提出并实现了分布多跳式网络吞吐量的分布式并行优化算法.首先将分布多跳式网络等效成M/M/m级联排队系统,并用流水线技术实现了优化算法.然后研究了用户移动速度和网络环境对吞吐量的影响,并以此得出一般的近似最优的分布式算法.最后分析了多用户之间的干扰问题对网络吞吐量的影响.仿真结果表明,并行优化算法可以提高分布多跳式网络的吞吐量和降低通信时延;理论分析结果也说明了在某些情况下可将干扰看作高斯噪声.

Traditional optimization methods of distributed network throughput don＇t meet the demand of the users in high mobility and high data rate .In order to improve the optimization property of distributed net-work throughput ,the parallel optimization algorithm in distributed network throughput optimization has been proposed and realized .Firstly ,the distributed network is equalized to an M/M/m queuing system , and the optimization algorithm is realized by pipeline techniques .T hen the impact of user velocity and net-work deployment on the throughput have been studied ,based on which the general optimal algorithm is obtained .Finally the effect of multi-user interference on the network throughput has been analyzed theo-retically .Simulation results show that the parallel optimization algorithm can improve the throughput and reduce communication delay of distributed networks .Theoretical analysis under interference circumstances verifies that interference can be viewed as Gaussian noise under certain cases .