摘要
针对可重构路由交换平台构件库中构件检索效率不高的现状,提出一种基于刻面的软硬件构件分类方法。基于该方法设计误差反向传播神经网络构件检索算法,分析表明该算法具有并行性强、效率高、速度快的特点,适合规模大、构成复杂的构件库检索问题。仿真结果证明,在构件数量较多的情况下,该算法性能优于传统检索算法。
Aiming at lower efficiency of components retrieval from component database in reconfigurable network routing platform,a component classification method of software and hardware based on facet is presented.Then a component retrieval algorithm based on Back Propagation Neutral Networks(BPNN) is designed.It is analyzed that the algorithm fits for component database with large size and complicated constitution because of its parallel execution,high efficiency and speed.It is indicated that the algorithm improves the retrieval performance in contrast with most of current retrieval algorithms by the emulation,especially on large size component database.
出处
《计算机工程》
CAS
CSCD
北大核心
2010年第19期277-279,282,共4页
Computer Engineering
基金
国家"863"计划基金资助项目(2008AA01A323)
关键词
可重构
路由
构件分类
构件检索
神经网络
reconfigurable
routing
component classification
component retrieval
neural network