An efficient ant colony algorithm for QoS anycast routing

Abstract

Anycast communication is a new Internet service defined in IPv6, and it can make a host communication with the one “nearest” member in a group of servers. The anycast routing problem with multiple QoS constrained is known to be NP-complete problem and we can't get satisfying results when using the precise method in polynomial time. In this paper, a QoS anycast routing algorithm based on ant colony algorithm is proposed. In order to search a reasonable anycast routing path, a felicitous estimate method of the optimization path is designed and the load balance is considered. By adjusting the pheromone in the iteration-best path and re-initializing the algorithm, this algorithm can efficiently avoid falling into local optimal path and enlarge global searching. The experimental results show that this algorithm can effectively find the global-best solution that satisfied the QoS constraint and balance the load of networks, and has better performance and efficiency.