摘要
针对计算机网络规模滞后、服务类型单一和服务质量没有保证等问题,将网络规划(网络改进)与流量工程作为有机整体为网络业务流量提供QoS服务,并在此基础上提出基于流量工程的新型网络改进算法.算法运用构造Harafy图和网络扩充启发式算法保证拓扑约束,基于多QoS约束路由算法满足流量工程约束;并使用改进遗传算法策略全局寻求改进费用最小网终及容量分配.仿真结果表明:算法实现最小化新增链路,使改进后拓扑仅略大于理想拓扑(Harary拓扑),88%的流量通过少跳数传输,且各链路带宽分配均匀(均方差σ=1.1).从而使算法在满足各网络约束的同时能够有效地均衡网络负载,避免链路拥塞,提高网络运行性能.
In order to solve the network＇ s service problems, such as narrowness of its scale, singleness of service types and no guarantee for quality of service （QoS） , network planning （network improvement） and traffic engineering（TE） were integrated together to respond to traffic＇ s QoS requirements. A network improvement algorithm based on TE was proposed. Constructing Harary topology and heuristic algorithm of network augmentation were used in the algorithm to satisfy the topological constraint, and the static routing algorithm based on multi-QoS requirements was used to satisfy the TE constraint. Then improved genetic algorithm was used to globally search the network of minimum improvement cost and its capacity of rational allocation. The simulation analyses show that the algorithm can minimize the incremental links, by which the improved topology is slightly bigger than the ideal one （ Harary topology）. Moreover, 88 % traffics are transmitted by few hops, and all link bandwidth is approximately uniform （ mean-variance σ = 1.1 ）. So, the network improvement algorithm can balance traffic, avoid link blocking and enhance the network behavior efficiently.
出处
《天津大学学报》
EI
CAS
CSCD
北大核心
2005年第10期865-868,共4页
Journal of Tianjin University(Science and Technology)
基金
教育部博士学科点基金资助项目(20030056007).
关键词
流量工程
网络改进
服务质量
遗传算法
可靠性
traffic engineering
network improvement
quality of service
genetic algorithm
reliability
