Optimizing active work: dynamical phase transitions, collective motion and jamming

Takahiro Nemoto, Étienne Fodor, Michael E. Cates, Robert L. Jack, Julien Tailleur

Published 2018-05-08Version 1

Active work measures how far the local self-forcing of active particles translates into real motion. Using Population Monte Carlo methods, we investigate large deviations in the active work for repulsive active Brownian disks. Minimizing the active work generically results in dynamical arrest; in contrast, despite the lack of aligning interactions, trajectories of high active work correspond to a collectively moving, aligned state. Conversely, we argue that alignment interactions in nature might be seen as an evolved strategy for maximizing active work. We use heuristic and analytic arguments to explain the origin of dynamical phase transitions separating the arrested, typical, and aligned regimes.