BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Europe/Stockholm
X-LIC-LOCATION:Europe/Stockholm
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260624T171340Z
LOCATION:Bldg. 8 - Entrance Hall
DTSTART;TZID=Europe/Stockholm:20260630T173000
DTEND;TZID=Europe/Stockholm:20260630T194500
UID:submissions.pasc-conference.org_PASC26_sess135_posC106@linklings.com
SUMMARY:ACMP03 - Deep Reinforcement Learning for Algorithm Selection in Mo
 lecular Dynamics Simulations
DESCRIPTION:Patrick Metscher (Technische Universität München)\n\nShort-ran
 ge particle simulations are crucial in physics and chemistry, requiring ef
 ficient neighbor search and interaction algorithms. The C++ library AutoPa
 s supports more than 100 algorithms, but no single algorithm is universall
 y optimal. Traditional manual tuning is impractical due to dynamic simulat
 ion conditions. To solve this problem, AutoPas implements automated tuning
  strategies such as full search, Predictive Tuning, and Reinforcement Lear
 ning. This work focuses on a Deep Reinforcement Learning (DRL) approach, w
 hich dynamically adapts to simulation conditions in two phases: exploratio
 n (benchmarking algorithms) and exploitation (predicting the best algorith
 m via a neural network). For the Meta-parameter tuning, simulation runtime
  results are cached, avoiding repeating expensive simulations. Evaluations
  show that DRL outperforms Predictive and Reinforcement Learning strategie
 s, meta-parameter tuning further enhances performance. Therefore, DRL offe
 rs a robust and adaptive solution for accelerating short-range particle si
 mulations.\n\n
END:VEVENT
END:VCALENDAR