To satisfy the necessity of a large bandwidth a design of Virtual Topologies is required in order to obtain the maximum performance of optical networks. The problem at hand, also known as Routing and Wavelength Assignation (RWA), is solved using a novel approach, based on multiobjective Ant Colony Optimization (ACO), proposing a conversion of the original RWA problem into a problem of traditional routing, modifying the graph that represents the optical network. For the resolution of the RWA problem, this work proposes new versions of the MultiObjective Ant Colony System - MOACS and the Multiobjective Max-Min Ant System - M3AS. Both ACO algorithms simultaneously minimize the total number hop count and the total number of wavelength switching for a set of given unicast demands. In this way, a set of optimal solutions, know as Pareto set, is calculated in only one run of the algorithms, without a priori restrictions.
|