QoS Routing Based On Clonal Computation

The rapid growth of multimedia application and the increasing demand on building value-added service generate more requests on predicable services over Internet. Quality of Service is a certain request for the development of network. There are two methods to solve the problem of QoS, the one is to control the nodes in network, the other is to control the whole network or the local network, that is traffic engineering.Based on this two methods and Clonal Computation, routing with imprecise information and traffic engineering in MPLS networks are researched in this dissertation. The major achievement of this dissertation is outlined in the following:The model of network with imprecise information and the corresponding solutions are researched in this dissertation. Candidate paths are found by the theory of probability under constraint of the hop counts, band width, delay and cost in our algorithm, then polyclonal algorithm is used to optimize the candidate paths to find a better multicast tree with imprecise information. It is shown that the algorithm has better performance than GA in the network with imprecise information in our experiment.The model of MPLS network with multicast is researched in this dissertation. Immune clonal algorithm for the multicast traffic engineering in MPLS networks is proposed in this dissertation. This algorithm is compared with GA and SPT in experiment and shown that it has better effectAn adaptive clonal algorithm for the traffic engineering in MPLS networks is proposed in this dissertation. In this algorithm, the traffic demand is split over multiple paths, and the maximum of link utilization is used for the optimized object in this algorithm. The algorithm is compared with GA and some traffic engineering algorithms under scale of network from 100 to 500 nodes with computer simulations. The simulation results have proved that the algorithm has faster speed and better performance in traffic engineering and load balancing.