An ACO algorithm to tackle aggregated multicast problem

Abstract

Traditional multicast technology faces a serious state scalability problem when there are large numbers of concurrent groups in the network. As a new approach to solve this scalability problem, aggregated multicast forces multiple multicast groups to share a common distribution tree. This can be defined as a minimum grouping problem and is proved to be an NPC problem. An ant colony optimization algorithm to tackle this problem is proposed. The number of groups in each aggregated tree is used as an important component when designing the fitness function between two multicast groups. Pheromone update rules are designed based on the fitness function. And the number of common neighbors between a multicast group and an aggregated tree is defined as the selection heuristic information. Simulation results show that this algorithm performs well with various bandwidth waste rates. Compared with a greedy algorithm, this algorithm has better optimization performance, especially when bandwidth waste rate is relatively big.