A modified ACO algorithm for multicast state scalability problem based on multicast tree similarity

Traditional IP multicast technology establishes and maintains a tree for each multicast group, and this leads to a serious state scalability problem when large numbers of multicast groups exist in a network. As a novel scheme to address this scalability problem, aggregated multicast forces multiple groups which have identical or similar trees to use one deliver tree. Given the number of multicast trees, searching for an aggregation solution to minimize the average bandwidth waste rate is an important optimization problem in aggregated multicast. In this paper, we propose a modified ant colony optimization algorithm for this optimization problem. We define the similarity of multicast trees, and combine it with similarity sequence information to define the selection heuristic information of our algorithm according to the characteristics of similarity. Simulation results indicate that our algorithm has better evolutionary ability, and the selection heuristic information can overcome the initial blindness and improve the convergence time of ACO algorithm. Compared with a greedy algorithm, our algorithm has better optimization performance.