NGI中的一种禁忌-递阶遗传QoS组播路由算法  

Tabu-Hierarchy Genetic QoS Multicast Routing Algorithm in NGI

作　　者：王兴伟[1] 侯美佳[1] 易秀双[2] 黄敏[1] 

机构地区：[1]东北大学信息科学与工程学院,辽宁沈阳110004 [2]东北大学计算中心,辽宁沈阳110004

出　　处：《东北大学学报（自然科学版）》2006年第5期493-496,共4页Journal of Northeastern University(Natural Science)

基　　金：新世纪优秀人才支持计划项目;国家自然科学基金资助项目(60473089);国家发改委CNGI示范工程项目(CNGI-04-13-2T;CNGI-04-6-2T;CNGI-04-15-7A)

摘　　要：设计了NGI主干网重要组网形式IP/DWDM光Internet中一种非NP类QoS(quality of service)组播路由算法,基于禁忌-递阶遗传算法(tabu-hierarchy genetic algorithm,THGA)构造优化的QoS组播路由树·该算法模拟生物繁衍过程,采用递阶编码方式,引入禁忌交叉和禁忌变异两种禁忌算子改进遗传操作,综合考虑了用户QoS需求和网络费用,同时有助于实现网络负载平衡·仿真结果表明,所设计的算法是可行和有效的,它在一定程度上克服了基于经典遗传算法的QoS组播路由算法存在的早熟和收敛慢等问题,提高了算法性能,较好地解决了IP/DWDM光Internet中的QoS组播路由问题·A non-NP QoS （quality of service） multieast routing algorithm in IP/DWDM optical lnternet, which is one of the main networking technique of NGI （next generation Internet） backbone, is proposed to construct an optimized QoS multieast routing tree based on THGA （tabu-hierarchy genetic algorithm）. The algorithm imitates the biologic propagation process and adopts the hierarchical coding scheme with two tabu operators （tabu crossover and tabu mutation） introduced to improve the genetic operation, taking account of user＇s requirements for QoS and network cost to benefit the balance in network load. Simulation results showed that the proposed algorithm is both feasible and effective in overcoming the specific shortcomings in the QoS multicast routing algorithm just based on the claxsical GA to a certain extent, such as prematurity and slow convergence. It not only improves the efficiency and performance of algorithm but provides a better way to construct the QoS multicast routing tree in IP/DWDM optical Internet.

关 键 词：下一代互联网(NGI) IP/DWDM光INTERNET 服务质量(QoS) 组播 路由 禁忌-递阶遗传算法 

分 类 号：TP393.14[自动化与计算机技术—计算机应用技术]