Multi-objective multicast routing optimization in Cognitive Radio Networks

In this paper, we study the multicast routing problem in Cognitive Radio Networks (CRNs). We propose a new network modeling method, where we model CRNs using a Multi-rate Multilayer Hyper-Graph (MMHG). Given a multicast session of the MMHG, our goal is to find the multicast routing trees that minimize the worst case end-to-end delay (delay), maximize the multicast rate (rate) and minimize the number of transmission links (numOfLinks) used in the multicast tree. We apply two metaheuristic algorithms (Multi-Objective Ant Colony System optimization algorithm (MOACS) [1] and A Simulated Annealing-Based Multi-objective Optimization Algorithm (AMOSA) [2]) in solving the problem. We also study the scheduling problem of multicast routing trees obtained using the MMHG model. Our simulation results show that within a few seconds, MOACS can find more than 60% of the approximated Pareto Front (APF) in small CRNs, and AMOSA can find approximately 45%. Moreover, the solutions found by MOACS and AMOSA that are not in the APF are within 10% relative distances to solutions in the APF.