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.
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.
出处
《华中理工大学学报》
CSCD
北大核心
1998年第10期57-59,共3页
Journal of Huazhong University of Science and Technology
基金
国家"九五"重点科技攻关资助项目(96-743-01-04-02)
关键词
输入缓存
神经网络
时延相关仲裁
交换网络
input-buffer
probability generation function
performance analysis
neural network
ar- bitration based on delay