BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Data-Driven and Equation-Informed Optimal Control of Lagrangian pa
 irs in turbulent flows - Chiara Calascibetta
DTSTART:20240226T130000Z
DTEND:20240226T140000Z
UID:TALK211642@talks.cam.ac.uk
CONTACT:Prof. John R. Taylor
DESCRIPTION:We show how to apply optimal control theory to catch a passive
  drifting target in a turbulent flow by\nan autonomous flowing agent with 
 limited maneuverability. For the case of a perfect knowledge of\nthe envir
 onment\, we show that Optimal Control theory can overcome chaotic dispersi
 on capturing\nthe Lagrangian target in the shortest possible time [1]. We 
 also provide baselines using heuristic\npolicies based on local-only hydro
 dynamical cues [2]. How to extend this approach to model-free\nReinforceme
 nt Learning tools is also briefly discussed [3]. Data are open downloadabl
 e from TURBLagr [4]\,\na database of more than 300K three-dimensional traj
 ectories of tracer particles advected\nby a fully developed homogeneous an
 d isotropic turbulent flow.\n\n(1) Calascibetta et al.\, Commun. Phys. 
 6\, 256 (2023).\n(2) Monthiller et al.\, Phys. Rev. Lett. 129\, 064502 
 (2022).\n(3) Calascibetta et al.\, Eur. Phys. J. E 46\, 9 (2023).\n(4)
  smart-turb.roma2.infn.it\n
LOCATION:MR5\, CMS
END:VEVENT
END:VCALENDAR