摘要
可重构路由器的研究与设计将冗长的网络设备生产过程变为类似于搭积木的构件装配过程。该文从构件组装角度出发,根据可重构路由器的特点,并针对第三方厂家大量提供构件导致的构件库急剧膨胀,构件提供困难的情况,提出了可重构路由器构件候选集CCS(Component Candidate Set)概念。该文针对构件静态加载和动态加载的区别,采用神经网络分别设计了静态构件候选集生成方法和动态候选集生成方法。CCS的提出解决了可重构路由器构件组装过程中的自动化和智能化的问题。CCS描述了构件运行状态、自主运行及自适应演化运行等机制,最后给出了基于该方法生成构件候选集的仿真实验,实验验证该候选集生成方法的准确性和高效性。通过该方法在可重构路由器中的实际运用证明该方法的实用性。
As a result of the research and design of Reconfiguration Router Unit （RRU）,the producing of network equipment is turned into component composition as building block.Based on the component composition,the character of RRU,and situation that the third-party manufacturers provide components for a large number of the component library led to the rapid expansion of the difficulties of the urgent need to provide components,this paper proposes the CCS （Component Candidate Set）.Based on the difference between the static load and dynamic load,the scheme of creating component candidate set is based on the neural network.The CSS is proposed to resolve the problem of automation and dynamic in the process of RRU component composition.According to the CCS,this paper describes the running state of component,self-determination running and adaptive evolution operation.Simulation experiment shows that this CCS-based method is precious and effective.The method that be successfully used in the reconfiguration router unit proves that the method is practical.
出处
《电子与信息学报》
EI
CSCD
北大核心
2010年第8期1980-1985,共6页
Journal of Electronics & Information Technology
基金
国家863计划项目(2008AA01A323)
国家973计划项目(2007CB307102)资助课题
关键词
可重构路由器
构件候选集
生成方法
神经网络
Reconfiguration Router Unit （RRU）
Component candidate set
Generation method
Neural network
