Genetic algorithms used for the optimization of light-emitting diodes and solar thermal collectors

We present a genetic algorithm (GA) we developed for the optimization of light-emitting diodes (LED) and solar thermal collectors. The surface of a LED can be covered by periodic structures whose geometrical and material parameters must be adjusted in order to maximize the extraction of light. The optimization of these parameters by the GA enabled us to get a light-extraction efficiency η of 11.0% from a GaN LED (for comparison, the flat material has a light-extraction efficiency η of only 3.7%). The solar thermal collector we considered consists of a waffle-shaped Al substrate with NiCrO_x and SnO₂ conformal coatings. We must in this case maximize the solar absorption α while minimizing the thermal emissivity ? in the infrared. A multi-objective genetic algorithm has to be implemented in this case in order to determine optimal geometrical parameters. The parameters we obtained using the multi-objective GA enable α~97.8% and ε~4.8%, which improves results achieved previously when considering a flat substrate. These two applications demonstrate the interest of genetic algorithms for addressing complex problems in physics.