An ant colony optimization algorithm to aggregated multicast using the idea of bin packing

Large-scale deployment of multicast applications is limited by the number of states that are set in routers for multicast groups. As a new approach to multicast state reduction, aggregated multicast forces multiple multicast groups sharing a common distribution tree. An ant colony optimization algorithm to aggregated multicast is proposed. Inspired by bin packing problem, relative fullness is used as an important component to define fitness function. To improve the algorithm's convergence time, heuristic information is introduced according to changes of aggregated trees' bandwidth waste rate. After each iteration a new pheromone update rule is proposed. Simulation results show that this algorithm performs well in scenarios with bigger bandwidth waste rate or larger network scale. Compared with greedy algorithm by running for the same amount of time and in the same network topology, the algorithm has better optimization performance.