Exploring the limits of multifunctionality in adaptable networks: comparing flow networks to mechanical metamaterials
ORAL
Abstract
Previous work shows that spring networks are both adaptable and robust - via selective bond pruning, specific functions can be programmed precisely, efficiently and robustly [Rocks et. al., 2016, arXiv:1607.08562]. These functions include localized but long-range-correlated deformations reminiscent of allostery in proteins. Analogous functionality can be introduced into flow networks by controlling the current through a bond in response to a current applied elsewhere in the network. Here we explore the limits of multifunctionality. How many separate independent functions can be simultaneously tuned successfully into a network and how many different targets can be controlled by a single source? These questions can be classified as constraint-satisfaction problems that we study in both mechanical and flow networks.
–