摘要
针对波分复用(WDM)光网络中动态选路和波长分配(RWA)问题,提出了一种基于改进蚁群算法的分布式动态RWA方法.在蚂蚁选路的概率中加入成本因素,并且只增加优秀路径上的信息素,从而对现有蚁群算法进行了改进,加快了其收敛速度.作者将改进的蚁群优化算法与分层图相结合,提出了一种构造时延受限的最小代价组播树的并行算法.仿真结果表明,与现有最短路经相比,该算法有效降低光路阻塞率,促进波长资源的合理分配,同时分布式的计算方法也降低了现代频繁变化的大型光网络的通信开销.
In this paper, the distributed method of dynamic routing and wavelength assignment(RWA) based on improved ant colony algorithm is brought forward in wavelength-division multiplexed optical networks. The paper presents an ameliorated ant colony optimization algorithm, which increase the speed of convergence by adding the cost factor to the routing probability and updating the pheromone only on the good path. On the basis of combining it with wavelength graphs, the authors propose a multicast routing algorithm using parallel computing to construct a minimal cost multicast tree under a given delay bound. Simulation result indicates that,compared with shortest algorithms, this algorithm can achieve lower blocking rates and rational assignment of wavelength resource, and decrease the communication overhead of current large networks which vary frequently.
出处
《四川大学学报(自然科学版)》
CAS
CSCD
北大核心
2008年第5期1062-1066,共5页
Journal of Sichuan University(Natural Science Edition)
关键词
WDM
RWA
波长分配
蚁群算法
阻塞率
RWA, wavelength assignment, ant colony algorithm, blocking rates