Skip to content

Latest commit

 

History

History
12 lines (9 loc) · 1.35 KB

README.md

File metadata and controls

12 lines (9 loc) · 1.35 KB
Raw

inverseLLE

Genetic-algorithm (GA)-based optimisation of resonator dispersion for Kerr comb state tailoring via the Lugiato-Lefever Equation (LLE). Check the Nature Photonics paper (DOI 10.1038/s41566-023-01252-7) or the arxiv version for more information.

Running the optimization

This is done by launching the script lle_dispersion_genetic_optimize.m. Two optimization scenarii are demonstrated:

  • Flat comb with a tageted power per line over a given bandwidth. This is currently the default scenario (see run init_polyCombTarget.m in the file). The initialization init_polyCombTarget.m can be modified to set a precise comb shape target and the fitness can be modified that the GA fits that target.
  • Octave-spaced dispersive wave uncomment the line % run init_octaveDW.m in lle_dispersion_genetic_optimize.m and the opmitization will seek to enhance the comb power in a specified comb line.

Analysing the optimization results

By default the genetic evolution results are stored in a .mat file in the Data folder. You can retrieve the optimum spectrum and dispersion profile by running analyze_genetic_dispersion_results.m. Additional options can be performed such as simulating a laser scan to check if the state can be reached by scanning the detuning across resonance.