摘要
为了提高电信网的呼叫接通率和平衡网络负荷 ,提出了全新的基于多 Agent系统和神经网络预测的智能化选路方法 .SUN工作站上的仿真表明 ,它良好的分布特性和智能决策能力使之比其他方法优越 ,并介绍了这种方法的软件仿真系统 ,包括类、Agent的设计 ,呼叫的生成方法和呼叫处理进程 ,以及消息机制 .
Routing is a key problem of the telecommunication networks and affects directly their switching rate and load balancing. The routing becomes more and more important as the telecommunications traffic increases at full speed. The results of demerit-analysis of routing methods being used in switching systems were applied to plan a kind of intelligent routing strategy which was based on multi-agent systems and neural network prediction. Simulations on SUN Workstation indicated it was better by virtue of its outstanding performance in optimizing load distribution through intelligent routing of messages. Its software design was introduced, including the design of class and agent, call generating and call transacting processes, and messages mechanism.
出处
《浙江大学学报（工学版）》
EI
CAS
CSCD
北大核心
2001年第6期592-597,共6页
Journal of Zhejiang University：Engineering Science
关键词
选路
AGENT
类
消息
Computer simulation
Computer software
Intelligent agents
Multi agent systems
Neural networks
Switching
Telecommunication traffic
