Abstract

Quality-of-service (QoS) based multicast routing is a major challenge to next generation networks due to the increasing demand of real-time applications which require strict QoS guarantee. In the presented multi-objective multicast routing, the QoS parameters, namely, cost and available bandwidth are represented as objectives, while end-to-end delay and delay jitter are represented as constraints. The optimization is strived using an elitist multi-objective evolutionary algorithm. The topological assisted tree structured encoding was proposed to represent the multicast tree. The individual solution or chromosome was represented as a combination of arrays where each array represents a random route from destination node in multicast group to source node. The effectiveness of the proposed algorithm is tested on various networks, including the network formed using network topology generator BRITE. The best compromise solution is obtained using fuzzy cardinal priority ranking. The performance of this algorithm was compared with weighted sum genetic algorithm. The simulation results demonstrate that the multi-objective optimization with the proposed encoding scheme is effective in providing faster and guaranteed convergence.

Key words: Multicast routing, multi-objective optimization, tree structured encoding, evolutionary algorithm, genetic algorithm.