A newly design method for designing multi-layered radiative cooling metamaterials based on evolutionary algorithms is exposed. The developed GA has been tested in three cases, resulting in three different structures that achieve, theoretically under direct sunlight, a net cooling power of 39.96 W/m 2 , 57.78 W/m 2 and 61.77 W/m 2 . Such devices are composed of 9, 15 and 24 layers respectively with a total thickness of less than 4.8 µm in the worst case. By the nature of the method, fewer design experience in metamaterials is needed, as well as it is freecost, due to the use of analytical calculations for the emissivity of the metamaterials instead of a commercial generic electromagnetic solver. Automated design of radiative cooling multi-layered structures and other applications in the infrared range can be further developed with this work.

In this work we present a novel way to design thinfilm radiative cooling metamaterials based on genetic algorithms. Three simulations with different design constraints have been done, resulting in three structures that achieve 39.96 W/m 2 , 57.78 W/m 2 and 61.77 W/m 2 under direct sunlight, respectively. These structures are shorter than 5 µm of height and are composed of 9, 15 and 24 layers. This design method has the advantages of being automatable, needs fewer design experience in metamaterials and does not rely on commercial simulators. This work opens the path to an easy way of automated design of thin-film multi-layered devices for radiative cooling and other applications in the infrared range.