Optimization of Passive Optical Networks by means of fiber nonlinearities interference reduction

In Passive Optical Network (PON), which usually uses intensity modulated signals to convey information fiber nonlinearities are enhanced due to high input and pump powers, short channel spacing, high bit rates, high number of channels, etc. This paper proposes an optimization algorithm applied to PON based on wavelength division multiplexing (WDM) ring. The optimization consists in setting up initial powers and frequencies, based on Multiobjective Genetic Algorithm (MGA) therefore fiber nonlinearities in the intensity modulated WDM channels are minimized. The MGA searches for the best channel spacing and initial powers so that the overall induced intensity distortion is minimized. The fiber nonlinearities are analyzed by the Volterra Series Transfer Function (VSTF) method that calculates the induced intensity distortion in each channel. The performance of the network is measured in terms of the eye diagram, which shows an improvement of 6 dB in the best optimization case.