| In computer networks, to provide multicast communication for multimedia real-time service, such as E- commerce, distance-education and videoconference, is one of the important issues in current research field. Multicast routing constructs paths along which data packets from a source are distributed to many, but not all, destinations in a communicational network. A fundamental issue in multicast is how to construct low-cost trees that satisfy the Quality of Service (QoS) requirements and spanning all destinations.The paper combined the method of immune clonal computing with Multi-Agent, and proposed two delay-constrained multicast algorithms. The main research work and results are follows:1. The existing heuristic algorithms for Steiner tree problem are researched;2. Principle and applications of Multi-Agent system were researched in this paper; and based on this, the paper also studied Multi-Agent Genetic Algorithm (MAGA);3. Immune operator is imported into the framework of MAGA, and aim at multicast routing problem, Multi-Agent multicast routing based on Immune Algorithm (MAIA) is proposed. Compare to traditional GA, MAIA uses transcendental knowledge of specific problems to direct and accelerate search. The algorithm solves multicast routing by combined with intelligent behaviors, such as competition, cooperation and self-study, of agent.4. Characteristics of Clonal Selection computation are combined with multi-agent, Multi-Agent multicast routing Based on Immune Clonal Selection Algorithm (MAICSA) is proposed in this paper. Firstly, for each member of multicast group, the algorithm gets the set of paths that satisfy the delay-constraint; and then, select candidates by using Clonal Strategies and Multi-Agent evolution to construct a multicast tree. It is shown that the algorithm has better global searching ability and effectively overcomes the drawbacks of GA with computer simulations. At the same time, the algorithm could be operated steadily, expediently and simply.
|
PDF Full Text Request