用非线性智能集成算法解决信息网拥塞问题  

Nonlinear Integrated Intelligent Algorithms Applied in Network Con gestion Control

作　　者：章敬东[1] 刘小辉[1] 邓飞其[1] 刘永清[1] 

机构地区：[1]华南理工大学自动化科学与工程学院,广州510641

出　　处：《计算机工程与应用》2003年第15期20-22,108,共4页Computer Engineering and Applications

基　　金：国家自然科学基金(编号:69934030);广东省计委高技术项目基金(编号:2001309)

摘　　要：该文将遗传算法和牛顿算法相结合,提出一种非线性智能集成算法解决信息网络拥塞问题,弥补了牛顿算法求率低、可靠性差以及遗传算法收敛速度慢的缺陷。用于网络拥塞控制的仿真结果表明,该算法能够高速可靠地拥塞模型的全局解,能有效解决网络拥塞问题,并使信元丢失率保持在CCITT要求的最优水平。Combining Genetic Algorithm with Newtonian Algorithm,this paper presents a kind of Nonlinear Integrated Intelligent Algorithm which can overcome disadvantages of low probability for global solution of Newtonian Algorithm and convergence tardiness of Genetic Algorithms to solve Network Congestion Problem.Computers group simulation results indicate that this new algorithm can efficiently solve Network Congestion problem and keep the leakage of information word on best level of CCITT criterion when people get fast convergent and reliable global solution.

关 键 词：智能集成 遗传算法 牛顿算法 非线性 网络拥塞控制 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]