The Research Of P2P Overlay Network Multicast Routing Intelligent Algorithm Based On Degree-Constraint

Recently, with the rapid development of computer and network technology, more and more multimedia services appear in the Internet, such as the live online、video conference、 remote teaching and so on. Such applications have high requirements for the network bandwidth and delay. Multicast has been considered as an effective solution. But IP multicast has great dependence on network equipment. So the equipment on the multicast path needs to be upgraded, which consumes a huge amount of money and hinders the deployment of IP multicast. This is also the reason for IP multicast not being widely used. With the urgent demand of network multicast, application layer multicast was proposed.At same time, P2P as a practical application of application layer is in constant development. There are many advantages in P2P network. So, the multicast problem in P2P network has become a hot topic in industry and academia filed. There are some typical P2P overlay network multicast, such as the Scribe, Chord multicast, CAN-Multicast and Bayeux. Owing to the rise of streaming media application use, amount of data needs to be transferred. The method of multicast or broadcast are used in P2P network for software update files, predicate the popular content. So every node is burdened with a heavy load for sending the data to its neighbor nodes. In order to reduce the pressure of the nodes and the links, it is effective to restrict the degree of the nodes. The application service servers expect to get a minimum cost solution. In short, data is transferred along a spanning tree of the P2P network, where the maximum vertex degree of the tree is limited to a constant and the cost of the tree is minimum. This solution is called as degree constraint minimum spanning tree in the graph theory.In this paper, the P2P application layer multicast was the research object. And we focus on how to efficiently find a degree constrained minimum multicast tree in the application layer multicast. The main work is as follows:(1) In this paper, a method based on ant colony algorithm with local search for degree constraint minimum spanning tree problem was proposed, and compared with other algorithms by simulation experiment. Ant colony algorithm finds the optimal solution by individual and collective cooperation. But it needs a lot of time to search the solution and is easy to fall into local optimum. So we proposed an algorithm based on the ant colony with local search which avoids the algorithm falling into a local optimum. The simulation results show that compared with other algorithms the new ant colony algorithm is better in the efficiency.(2) The efficiency of the algorithm is improved a lot after we changed the ant colony algorithm. We also research the estimation of distribution algorithm (EDA) which has the property of fast convergence and proposed the new method of based on the EDA with local search algorithm. The simulation results show that the new algorithm is better than the previous algorithm in convergence time which means this method is feasible.