Genetic Approach for Optimizing the Placement of All-Optical Regenerators in WSON

Abstract

In wavelength switched optical networks (WSONs), the placement of opto-electronic regenerators has always been optimized to keep the cost and power consumption contained, while ensuring the quality of transmission (QoT) for the optical signals. Nowadays, with the increase of the transmission rates and the spectral efficiency of the modulation formats, optical signal regeneration is still relevant for ensuring QoT. The progress in integrated photonics makes possible the realization of integrated optical regenerators that can replace the traditional opto-electronic regenerators. All-optical regenerators enable a reduction of size and energy consumption. However, differently from opto-electronic regenerators, recently fabricated integrated optical regenerators neither perform wavelength conversion nor regenerate multiple lightpaths at the same wavelength. Thus, novel constraints and limitations are introduced to the conventional regenerator placement problem, which has been well studied in the past for opto-electronic regenerators. This paper addresses the all-optical regenerator placement (ORP) problem in WSONs with guaranteed QoT. A genetic algorithm (GA) is proposed for optimizing the ORP while jointly solving the routing and wavelength assignment problem. GA results are compared with the optimal solutions found by solving a binary linear programming (BLP) formulation. When properly tuned, GA is able to quickly achieve the optimal solutions found by the BLP solver. The performance analysis indicates that the additional constraints have a negligible impact on the overall number of regenerators, compared to the case with opto-electronic regenerators.