Addressing the GRWA problem in WDM networks with a tabu search algorithm

The traffic grooming, routing and wavelength assignment (GRWA) problem in wavelength division multiplexed networks is addressed. A new heuristic algorithm, the GRWABOU algorithm, has been developed from tabu search techniques. Its objective function seeks the minimization of network costs through the minimization of the number of electronic cards used in network nodes. It uses an evaluation function that, in addition to depending on the objective function, provides means to handle constraint violations in order to travel in the non-feasible domain in a controlled way. It can be applied to ring or general mesh networks. It accepts any type of traffic demand with any type of granularity on the client side. On the transport side, it supports multiple bit rates simultaneously. Computational experiments on three different mesh networks are presented, showing the versatility of the algorithm not only for cost minimization studies but also for other types of studies concerning, as example, number of hops, wavelength capacity usage and multi-service provisioning platform (MSPP) port usage