Genetic algorithm based single pulse energy optimization in Mamyshev oscillator

The pulse energy in mode-locking fiber lasers is fundamentally limited by non-linear phase accumulation. However, the exploitation of this accumulated phase to regenerate the spectrum was successful of increasing the energy per pulse in Mamyshev oscillator. In such a configuration, whether experimentally or numerically, it is almost impossible to find the exact cavity parameters that maximize the energy. In that context, we propose here for the first time a very efficient method based on the genetic algorithm for selecting the tunable filter parameters giving the high pulse energy in the single pulse regime. To this end, five key parameters of the laser system are optimized: the bandwidths of two filters, their separation, and the orders of the two super-Gaussian filters. The genetic algorithm is meticulously tailored to explore the complex parameter space, with the fitness function designed to maximize the single pulse energy while maintaining stability in the laser operation. It is demonstrated that, for a given small signal gain and fixed cavity parameters a specific relationship between filter bandwidths and filter separation must be achieved to obtain high pulse-energy performance. Numerical simulations have demonstrated the generation of stable pulses with a duration of 760 fs, energy of 1.49 nJ, and a peak power reaching 1.5 kW.