Reliability Optimization of Overlay Multicast Topology with Efficient Genetic Algorithm

Genetic algorithm (GA) has been widely used in a variety of communication network design problems. The encoding method is the essential issue for GA. In this paper we investigate the topology optimization of the tree-like overlay multicast systems with degree constraints. A new encoding method based on the degree-weighted array is proposed. It requires no repair strategies for the encoding, crossover and mutation operators. Compared with other topology optimization methods, the proposed approach considerably enhances the reliability of the overlay systems and decreases the transmission delay from the source node to the end node. The population size is an important design parameter of GA but it has little effect on the performance of the proposed approach in terms of reliability and transmission delay. The required CPU time depends on network size and population size, varying in the range from hundreds of milliseconds to tens seconds. Therefore the proposed GA is effective and efficient to improve the performance of a realistic overlay multicast.