Traffic Grooming in Multi-Layer WDM Networks: Meta-Heuristics versus Sequential Algorithms

This paper deals with electrical traffic grooming into optical lightpaths. In the context of a multi-layer EXC/OXC network, electrical traffic grooming aims to optimize resources utilization in terms of required optical channels. It also enables to relax the wavelength continuity constraint. Whereas most of the existing studies are applied either to permanent traffic demands or to dynamic traffic demands, we consider dynamic scheduled traffic demands for which the date of activation and life duration are pre-known. Our objective is to compare two algorithmic approaches: Meta-heuristics and sequential algorithms. The former approach deals with the scheduled demands as a set. It performs global optimization by considering all the possible grooming combinations. The latter approach tries to route the demands iteratively at their instant of arrival. We compare our algorithmic approaches with two other grooming strategies proposed in the literature.