Intelligent and fast IRWA algorithm based on power series and Particle Swarm Optimization

In all-optical networks signals are transmitted through physical layer with no regeneration. Therefore, physical impairments along lightpath can severely reduce network performance. For this reason, many efforts have been made to develop impairment aware routing and wavelength assignment algorithms (IRWA) in order to mitigate the impairments effects, improving the network performance. In this paper we propose and analyze the performance of an adaptive impairment aware routing algorithm based on a set of chosen input network parameters. The cost function of this routing algorithm is based on a power series expansion. The routing algorithm, called Power Series Routing (PSR), is trained by an optimization technique called Particle Swarm Optimization. We show that this IRWA algorithm can learn during the training stage and adapt itself to the network conditions.