一种基于优化思想的网络拥塞控制算法

分析了现有网络优化框架的基本模型,研究了更一般的条件,将静态模型扩展到了动态模型,使模型的应用范围更广。运用子梯度方法求解新模型的对偶问题,得到了一种新的基于对偶问题的分布式算法框架,并证明了这种新算法的收敛性。同时,又证明了已有的算法只是新算法的一个特例。因而本文给出了一类具有更一般意义的分布式算法框架,推广了已有的网络模型算法。

An optimization-based internet networks congestion control algorithm

In this paper,we describe an optimization-based framework for internet and propose a dual model for end-to-end congestion control.Through analyzing the characteristic of the source algorithms and AQM at the equilibrium,we propose a new model with less constraint condition and a distribution algorithm with larger application.Simulation studies prove the stability and convergence.The theory explain the reason why current algorithm can converge although the change of link capacity.