Research On Multiple Constrained QoS Multicast Routing Optimization Algorithm

With the rapid development of network technologies, video conferencing, distance education, online games and other real-time business make the rise of the network bearing capacity of a higher demand, so multicast communication method to become the main form of multimedia services. Real-time services constraint different QoS needs (delay, jitter, bandwidth, packet loss rate, etc.), and the requirements of QoS multicast routing algorithm is also relatively high, the research of multiple multicast routing with QoS constrained algorithm has become a hotspot issue in today’s network. QoS multicast routing problem is to find one rout which meet the QoS constraints of the optimal multicast tree with minimum cost. This issue has already been shown to be NP problem, and multicast routing problem constrained multiple QoS is a NPC problem which has no solution made by an algorithm of the polynomial level, so many intelligent algorithm applied to this area.Genetic algorithm is the main intelligent algorithm used in this paper to solve the problems of QoS multicast routing algorithm , genetic algorithm is easy to fall into premature convergence, so we use biological immunity to suppress the local optimum and premature convergence. On this basis, use a sub-optimal mechanism to determine whether the current population into local optimum, if Algorithm stick to local optimal,and the immune algorithm is used to suppress the premature convergence and out of the local optimum.On this basis, proposed a load balancing based on immune genetic algorithm for QoS multicast routing, The traditional QoS multicast routing problem is usually better although the multicast problem, but the cost of a small path is always used frequently, which may cause network congestion, and the resource utilization is relatively low. The LB-IGAE algorithm proposed by this paper simultaneously concerns the load of the path and the load of the node, the utilization rate of the path and the degree constraint function are also added in the fitness function of the genetic algorithm. The simulated experiments indicate that this algorithm can find the multicast tree with the minimum cost, and it can make the links and the nodes have the better performance of load balance with satifying the QoS demand. The algorithm has better comprehensive performance.