| With the rapid development of the internet, some multimedia services, such as video ondemand and remote education, have become reality in our daily life. These newservices based on multicast routing technology, and have strict demands on the quality ofservice(QoS) controlling according to the service itself and the user demand. Like bandwidthand latency requirements, picture distortion or audio-video synchronization during datatransmission. However, traditional best-effort routing technology can not ensure quality ofservice, which make QoS multicast routing become a more and more important subject for theresearchers. The key of QoS multicast problem is routing selection, that is finding the pathsfrom the source to the destinattion nodes with minimal cost while satisfying the QoSconstraints, according to the current network status and topology. This is a NPC problem,which is difficult to solve, especially with multiple QoS constrains. GeneticAlgorithm (GA) and artificial neural network (ANN) can be used to solve the complexcombinational optimization problem, thus using the two intelligent algorithm to solve theQoS multicast routing problem is an idea of great concern.In this article, we proposed two new algorithms to solve the QoS multicast routingproblem. One is baesd on GA, the multicast tree coresponding to the solution is coded as thechromosome, and only the upper triangular adacency matrix of the network nodes is used,therefore the code length is decreased. In addition, elitist strategy is used to boost theconvergent of this algorithm. The other is based on ANN, and use the GA to do parameteroptimization. Thus this algorithm takes andvance on the strong global random searchcapability of GA, as well as the robustness and self-learning ability of neural networks. Thesimulation experiment shows that, the two algorithms has better performance in solving themulti-constraints QoS multicast routing problem comparing with the-only GA/ANN methods. |
PDF Full Text Request