Abstract
For the reliable real time video transmission of overlay network, optimal multicast service nodes (MSNs) should be selected to build highly efficient hierarchical overlay multicast tree. In this paper, a MSNs selection algorithm based on immune evolution is proposed. MSNs are encoded by real-coded mechanism, and the K-medoids clustering distance is used to measure the similarity between MSN and other nodes. In accordance with the actual characteristics of MSN, a fitness function including penalty factor is brought forward to prevent undesirable individuals participating in evolution. Also, on the base of SGA, immune vaccines are extracted and imported into the evolutionary population to improve the global optimizing capability of algorithm. Theoretical analysis and simulation results show that this algorithm is effective and feasible. As to select MSNs, it not only overcomes the imitation-sensitive problem and local convergence caused by the traditional K-medoids algorithm but also avoid degradation phenomenon appeared in standard genetic algorithm , and improve search capabilities and speed convergence.