Optimization of Supercontinuum Spectrum Using Genetic Algorithms on Service-Oriented Grids

This paper describes the developments made towards the optimization of photonic crystal fibers using automated procedures based on Genetic Algorithms. The large amount of computations involved during the optimization process demand the usage of Grid technologies that are able to cope with these computational requirements. This paper describes the integration of a service-oriented Grid metascheduler and a Genetic Algorithm in order to create an automated tool which is able to optimize this kind of fibers. This work also provides an evaluation of these computational techniques for the optimization of the output supercontinuum (SC) spectrum generated in an optical fiber of a fixed length and fixed dispersion properties. The developed tool provides an scalable solution to this computationally intensive problem, and enable the optimization of other more complex designs.