Abstract: This paper studies the dynamic traffic grooming problem in IP/DWDM optical Internet，which has been proved a NP-hard problem.Heuristic and intelligent algorithms are adopted to solve the problem.In this paper，network and mathematics models are set up，and an artificial immune algorithm is designed to route and assign wavelengths for the new-coming connection request，minimizing the cost for the request as well.In order to check the feasibility and the validity of the algorithms，a simulation environment is developed with VC++ 6.0 and several practical optical networks，such as the NSFNET，the CERNET1，the CERNET2 and the GéANT，are tested.The experimental results show that the algorithm presented here can gain better result than the current heuristic algorithm.

摘要： 研究IP/DWDM光Internet中网状拓扑下的动态通信量疏导问题。网状拓扑下的通信量疏导问题已证明是NP难问题，需要采用启发式算法或智能优化算法来解决。针对动态通信量疏导问题，建立了网络和数学模型，提出了求解该问题的人工免疫算法，为新到达的通信量请求路由和分配带宽等网络资源，同时最小化满足该通信量请求的网络费用。为验证算法的可行性和有效性，用VC++6.0开发了一个仿真环境，同时以美国自然科学基金网NSFNET、中国教育和科研计算机网CERNET1和CERNET2以及欧洲巨人网GéANT等的骨干网拓扑为仿真用实例，与现有启发式算法进行性能比较，并对网络参数对算法的影响进行了分析。实验表明，提出的算法可以获取比现有启发式算法更加优化的解。