Modified GA with Adaptive Crossover and Mutation for RWA in Large-Scale Optical Backbone Networks

Abstract

The Routing and Wavelength Assignment (RWA) is a fundamental challenge in wavelength-routed all-optical networks, directly impacting network cost, resource utilization, and blocking probability. For a given network topology, RWA seeks to identify optimal lightpaths and assign wavelengths such that no two connections sharing a link use the same wavelength. Due to its NP-complete nature, exact solutions are impractical for large-scale networks, motivating the use of metaheuristic approaches. This work proposes a Genetic Algorithm (GA) based RWA solution incorporating node-connectivity-aware crossover and mutation strategies. A novel tournament selection method is introduced, combining tournament scores with the fitness function to achieve an effective balance between exploration and exploitation. The objective function simultaneously minimizes wavelength usage and the number of hops per lightpath, thereby reducing overall blocking probability. Simulation results demonstrate that the proposed GA approach outperforms conventional GA methods, achieving improved resource efficiency and reduced blocking in diverse network scenarios.

Keywords