978-1-4244-2626-3/08/$25.00 ©2008 IEEE On the Use of Genetic Algorithms for Solving the RWA Problem Employing the Maximum Quantity of Edge Disjoint Paths

In this paper, the routing and wavelength assignment problem in all optical networks is considered. A solution to the RWA problem is proposed through a Genetic Algorithm (GA) based approach to search for the most suitable set of paths from many possible ones. We use GAs to solve this problem by introducing a new fitness function based on the mean value of occurrence of the common edges between the different genes of a single chromosome. In addition, we further improve performance and particularly execution times by considering the maximum quantity of edge disjoint paths (EDPs) to initialize the population with some indigo chromosomes. These EDPs are pre-computed and inserted in the GA-RWA problem. The proposed approach has been evaluated on the NSF network topology and compared with other similar schemes. 1