Online Design

Abstract

In this chapter, we explore the possibility of building optimization algorithms for the specific purpose of solving materials design problems by integrating them with statistical physics. Specifically, we construct a formalism that can be used to turn statistical physics models that describe materials into optimizers that tailor them. On two simple test problems, we show that the resulting algorithms can be fast and efficient: we use our framework to trap a particle randomly walking on a substrate and to find optimal interaction energies that cause a simply polymer model to fold into a target shape. The speed of our approach suggests that optimizers developed by our new formalism might fill a niche in tailoring materials where computational expensive simulations prohibit alternative, established methods.