Multi Criterion Multicast Routing Algorithm Comparison for Large Networks

Abstract

Abstract QoS enabled multicast routing is known to be of non-polynominal complexity, which leads to the necessity of using heuristic algorithms to find sub-optimal solutions to the problems of this class. The evaluation of such algorithms requires the use of the simulation techniques as the heuristics’ results are of stochastic nature. Because of the problem complexity the simulation times increase significantly in the function of the network size, therefore the results presented in the literature are often limited to only small models. In this article the results of the evaluation of different multicast QoS routing algorithms (further referred to as the Multi-Constrained Minimum Steiner Tree Problem - MCMST) have been presented for a wide range of network sizes reaching thousands of nodes.