首页 | 本学科首页   官方微博 | 高级检索  
     

基于时延策略的输入缓存交换网络仲裁与分析
引用本文:喻莉,石冰心,李佳,张斌. 基于时延策略的输入缓存交换网络仲裁与分析[J]. 华中科技大学学报(自然科学版), 1998, 0(10)
作者姓名:喻莉  石冰心  李佳  张斌
作者单位:华中理工大学电子与信息工程系
基金项目:国家“九五”重点科技攻关资助项目(96-743-01-04-02)
摘    要:借助概率母函数的分析方法,对输入缓存方式交换网络的性能进行了理论分析,扩展了前人的分析结果,推出了吞吐率和延时表达式的闭和形式;并对这种网络的输出竞争,提出一种简单的基于时延相关策略的神经网络仲裁算法,解决信元调度问题.仿真结果表明,该仲裁方法改善了网络性能.

关 键 词:输入缓存  概率母函数  性能分析  神经网络  时延相关仲裁

Performance Analysis and Arbitration Algorithm Based on Delay of the Input-Buffered Switch
Yu Li,Doctoral Candidate, Dept. of Electronics , Information Eng.,HUST,Wuhan ,China. Shi Bingxin,Li Jia,Zhang Bin. Performance Analysis and Arbitration Algorithm Based on Delay of the Input-Buffered Switch[J]. JOURNAL OF HUAZHONG UNIVERSITY OF SCIENCE AND TECHNOLOGY.NATURE SCIENCE, 1998, 0(10)
Authors:Yu Li  Doctoral Candidate   Dept. of Electronics & Information Eng.  HUST  Wuhan   China. Shi Bingxin  Li Jia  Zhang Bin
Affiliation:Yu Li,Doctoral Candidate, Dept. of Electronics & Information Eng.,HUST,Wuhan 430074,China. Shi Bingxin,Li Jia,Zhang Bin
Abstract:The performances of the input-buffered switch fabrics are analyzed by means of the proba- bility generation function method. A simple neural network arbitration algorithm based on delay is proposed in order to schedule the head of line of the switch. The algorithm is fast and very easy to im- plement. The simulation results show that the performances are improved. Therefore, the method is effective and realistic.
Keywords:input-buffer  probability generation function  performance analysis  neural network  ar- bitration based on delay
本文献已被 CNKI 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号