Abstract
This paper deals with the offline allocation of lightpaths in all-optical networks taking into consideration not only the wavelength continuity and non-overlapping constraints of the classic routing and wavelength assignment (RWA) problem but also the optical nonlinear impairments modelled by the Gaussian Noise (GN) model. In this work, we deal with the associated impairment aware RWA problem aiming to allocate lightpaths for all origin-destination node pairs following the max-min fairness principle: we aim to maximize the number of lightpaths of the node pairs that can have less number of lightpaths. The paper presents integer linear programming methods, one exact method and two heuristic methods, based on path formulations to solve the problem. Both heuristic methods can be tuned to provide different tradeoffs between runtime and solution quality. Computational results are presented to assess the merits of the heuristic methods when compared with the exact method.