We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:

References & Citations


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Condensed Matter > Statistical Mechanics

Title: Probing the Theoretical and Computational Limits of Dissipative Design

Abstract: Self-assembly, the process by which interacting components form well-defined and often intricate structures, is typically thought of as a spontaneous process arising from equilibrium dynamics. When a system is driven by external \emph{nonequilibrium} forces, states statistically inaccessible to the equilibrium dynamics can arise, a process sometimes termed direct self-assembly. However, if we fix a given target state and a set of external control variables, it is not well-understood i) how to design a protocol to drive the system towards the desired state nor ii) the energetic cost of persistently perturbing the stationary distribution. Here we derive a bound that relates the proximity to the chosen target with the dissipation associated with the external drive, showing that high-dimensional external control can guide systems towards target distribution but with an inevitable entropic cost. Remarkably, the bound holds arbitrarily far from equilibrium. Secondly, we investigate the performance of deep reinforcement learning algorithms and provide evidence for the realizability of complex protocols that stabilize otherwise inaccessible states of matter.
Subjects: Statistical Mechanics (cond-mat.stat-mech)
DOI: 10.1063/5.0067695
Cite as: arXiv:2108.05452 [cond-mat.stat-mech]
  (or arXiv:2108.05452v1 [cond-mat.stat-mech] for this version)

Submission history

From: Grant Rotskoff [view email]
[v1] Wed, 11 Aug 2021 21:20:14 GMT (26398kb,D)

Link back to: arXiv, form interface, contact.