Design of a flat-gain multipumped distributed fiber Raman amplifier by particle swarm optimization.

The pumping scheme of multipumped distributed fiber Raman amplifiers is optimized by a powerful method called particle swarm optimization. By use of particle swarm optimization, we optimize both pump powers and frequencies of multipumped Raman amplifiers with a high number of pumps. Particle swarm optimization is a fast and effective method, and it surpasses other optimization methods, such as the genetic algorithm, for optimizing fiber amplifiers. It is shown that the computational efficiency of particle swarm optimization is significantly better than that of the genetic algorithm, reducing the time of computation to one third, and its implementation is more straightforward. A gain bandwidth of 92.1 nm and a gain variation of 0.49 dB in the range of 1524.5-1616.6 nm are obtained by this method, using ten backward pumps in a 60-km-long amplifier. The gain variation reduction is due to the inclusion of pump frequencies in the optimization process.