Metamaterials are not found in nature; they are meticulously designed by scientists and engineers. Through their expert knowledge of simple mechanisms, they are able to rationally design complex geometries capable of rare, counterintuitive, and hopefully useful properties. In this research group, we ask whether we can move beyond rational design and use computational techniques to improve the design and functionality of mechanical metamaterials.
We do this in two ways. On the one hand we develop computational tools with which we try to unravel the complex relation between design and functionality in a multifunctional metamaterial. On the other hand we employ machine learning to learn such complex relations from data. In this way, we hope to create new, more complex designs with increased functionality, and to move beyond rational design.
Ryan van Mastrigt, David Dykstra