The Research Of Multicast Traffics Based On Genetic Algorithm In Sparse-Splitting Optical WDM Networks

Introducing multicast technique in optical layer can bring many valuable wide-band real-time applications, such as video distribution, teleconferencing, distant medical treatment, etc., But the implementation of all-optical multicast performance asks the nodes in the optical networks to be provided with the light-splitting capability. The node with such capability can forward an incoming optical message to more than one output link. The introduction of MC (Multicast Capable) node doesn't need to separate the multicast traffic in optical layer into many unicast traffics so as to save optical-path resources. However the cost of MC node is very expensive, so an effective method is that only a subset of the nodes are assumed to be split capable nodes, which is called sparse-splitting.At present the multicast research in sparse-splitting optical network is emphasized on the single traffic, the situation of existing multi-traffics has gained less research. In this dissertation a MC-RWA algorithm suited in sparse-splitting optical networks is first constructed, compared with the previous light-tree establishment scheme based on the source node, we adopt the scheme based on MC node. At the same time we introduce the genetic algorithm to collocate the multicast traffics, then construct a heuristic algorithm to optimize the number of wavelength and link used in the networks. About the genetic algorithm we have proposed a new coding and decoding scheme, and compared the effect of a variety of crossover and mutation operators with numerical simulation. Based on the above work the improved schemes of bidirectional decoding and...