Design of nondeterministic architected structures via bioinspired distributed agents.

Abstract

Nature manufactures structures via decentralized processes involving groups of agents. This is fundamentally different from traditional manufacturing, where objects are produced via sequences of predefined steps. In this work, we explore the idea of using simulated "swarms" of simple agents to generate new designs for architected materials in a decentralized, bioinspired manner. Individual agents choose their own actions based solely on information in their immediate environment, with no centralized control. The structures that these processes produce are the result of the collective action of the individual agents, rather than a predetermined design. We build an integrated platform for determining "rule-structure-property" relationships, analogous to process-structure-property relationships in materials science. The platform simulates agent behaviors to show how different rules and different environments result in different structures. We then three-dimensional print these and perform finite element analysis to experimentally and numerically characterize mechanical properties, including tensile strength and energy dissipation.