Optimization of frequency combs spectral-flatness using evolutionary algorithm

Abstract

We demonstrate the use of meta-heuristics algorithms for flatness optimization of optical frequency combs (OFCs). Without any additional component for flatness compensation, the laser alone is explored when driven by optimized bias current and radio frequency (RF) driving signals composed by multiple harmonics. The bias current amplitude and RF harmonic amplitudes and relative phases are optimized using particle swarm optimization (PSO) and differential evolution (DE) algorithms. The numerical results lead to a 9 lines-GS-laser-based OFC spectrum with 2.9 dB flatness. An online experimental optimization using the DE algorithm results in a 7-line-GS-laser-based OFC with 2 dB flatness.

Optical Frequency Comb Generation and Further Advances From the Photonics Community in Serbia and Western Balkans

“…The research conducted at the University of Belgrade-School of Electrical Engineering targets optical frequency combs for applications in both optical communications, i.e., combs comprising teeth separation in order of GHz [22], [23], and spectroscopy, requiring order of MHz teeth separation, or dual combs with heterodyne frequencies in the order of kHz [24]. In both cases the research aims to simplify the comb generation scheme exploiting the semiconductor laser gain switching technique with a goal of producing high bandwidth combs with better line flatness.…”

“…Our approach tends to simplify the scheme to a gain-switched laser and one intensity modulator with optimized performance biased with only 1.5 V peak-topeak RF signal to yield an OFC with 12 comb lines with 9.5 GHz separation [22]. Further simplifications include applying specific optimization techniques to maximize the number of comb lines generated directly from the gain-switched laser where we obtain 7 comb lines with 5 GHz separation [23]. The similar technique has been very recently applied to a scheme with external modulators cascade to obtain 15 lines with 10 GHz spacing [27].…”

Online versus Offline Optimization Methods for Raman Amplifier Optimization

“…The DE algorithm is a derivative-free, population-driven, and bio-inspired optimization method [15] [17]. It relies on iteratively going through three stages: mutation, crossover, and fitness evaluation.…”

“…However, ES can also be applied online, directly interacting with the experimental setup to evaluate the solutions. Some examples are the optimization of frequency combs [15] and optical transmitters [16]. Nevertheless, online ES approaches have not yet been applied to the RA case.…”

Inverse design of programmable optical frequency comb using deep learning

“…The OFC with 7 comb lines and good flatness (<2 dB) is obtained. Pinto et al also use a differential evolution algorithm to optimize the amplitude and phase of multiple RF signals to modulate the laser and obtain 7 comb lines with a flatness of about 2 dB in the experiment [24]. Although the optimization algorithm can search a large variable space more easily and get a global optimal solution, it is disadvantages are obvious: the complexity of the algorithm will rise sharply with the increase of parameters number, and it becomes easy to fall into local optimal solution.…”