Complex Systems Seminar: 'Low rattling: A predictive principle for self-organization in active collectives'

VIRTUAL SEMINAR LINK: https://umich.zoom.us/j/96616169868

ABSTRACT: In this work we suggest a mechanism for self-organization of active matter, which we believe may be quite general. This mechanism is similar in spirit to thermophoretic drift in colloidal suspensions, where particles gravitate to low-temperature regions. The difference is that here instead of suspended particles in a bath, we think of the dynamics of a complex system's state in its full high-dimensional configuration space. The temperature landscape is then replaced by what we call "Rattling landscape," which reflects how different system states respond to the driving forces. This way the system gravitates towards configurations that have special response properties to the external forces, giving the impression that it "adapts to its environment." As a proof of principle, we use our theory to predict and control the behavior of a simple robotic swarm.