基于模糊神经网络及Agent控制器的网络路由数值调整研究

A Study on the Dynamic Adjusting of Network Weights by Fuzzy Neural Network Agent Controller

基于典型多输入、单输出模糊控制系统,建立了基于模糊控制系统的神经网络(FNN)模型,并根据Agent网络控制系统的特点设计了基于模糊神经网(FNN)的Agent控制器,对Agent网络管理系统的网络权值进行了动态调整研究.研究结果表明,采用基于FNN的Agent控制器可对Agent网管系统的网络权值进行动态调整,进而可实现大型复杂网络系统中A-gent的最优动态路由序列的计算.

Based on the typical fuzzy control system with multi - input and single output, one neural network model has been established. One Agent controller based on the fuzzy neural network （FNN） has been designed in terms of the features of the Agent network control system, and the discussion on adjusting the network weights in Agent Network Management System dynamically has produced the results that show the network weights of Agent Net- work Management System can be adjusted dynamically by FNN agent controller. Moreover, the optimal dynamic route sequence of Agent in a large-scale complex network system can be calculated by FNN agent controller.