Abstract

While the Routing and Wavelength Assignment (RWA) problem has been widely studied, very few studies attempt to solve realistic size instances, namely, with 100 wavelengths per fiber and a few hundred nodes. Indeed, state of the art is closer to around 20 nodes and 30 wavelengths. In this study, we are interested in reducing the gap between realistic data sets and testbed instances, using exact methods.

We propose different algorithms that lead to solve exactly or near exactly much larger instances than in the literature, with up to 150 wavelengths and 90 nodes. Extensive numerical experiences are conducted on both the static and the dynamic cases. For the latter, we investigate how much bandwidth is wasted when no lightpath re-arrangement is allowed, and compare it with the number of lightpath re-arrangement it requires in order to fully maximize the grade of service. Results show that the amount of lightpath re-arrangement remains very small in comparison to the amount of wasted bandwidth if not done.

The Routing and Spectrum Assignment (RSA) problem is a much more difficult problem than RWA, considered in elastic optical networks. Although investigated extensively, there is still a gap between the size of the instances that can be solved using the current heuristic or exact algorithms, and the size of the instances arising in the industry. As the second objective of this study, we aim to reduce the gap between the two, using a new mathematical modeling, and compare its performance with the best previous algorithms/models on realistic data instances.