Fig. 2: Convergence of the genetic algorithm towards target microcomb.

a Convergence plots for three separate realizations of the genetic algorithm (GA) that show the fitness score improvement as the generation number increases, eventually converging to better values. Dotted lines depict the average fitness score corresponding to each realization of the GA, while solid lines represent the best score within each generation. The line colors—blue, green, and purple—correspond to the microcombs displayed in b–d, respectively. b–d Spectral output at the labeled points of each convergence plot. Each drop in the convergence plots in a shows a progress in refining the output microcombs (b–d), enabling us to observe the improvement in intermediate microcomb states, as we approach the desired target. The GA realization corresponding to c was initiated with a better gene pool compared to b and d. Hence, the corresponding average fitness plot in a achieves overall lower values compared to the others.