Computational Materials Discovery Using Evolutionary Algorithms
Evolutionary algorithms, as implemented in the USPEX code, offer a very attractive and powerful way for computational materials discovery, and this is demonstrated by their very wide applications. In this chapter we review this methodology, its adaptation to variable-composition systems, surfaces, nanoparticles, molecular crystals, and proteins. We discuss the reasons why it works so well. We discuss a number of interesting applications of this methodology to materials discovery—including superhard materials, high-Tc superconductors, high energy density materials, electrides, 2D-materials, polymers, and chemically anomalous solids and nanoparticles.