BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Los_Angeles
X-LIC-LOCATION:America/Los_Angeles
BEGIN:DAYLIGHT
TZOFFSETFROM:-0800
TZOFFSETTO:-0700
TZNAME:PDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0700
TZOFFSETTO:-0800
TZNAME:PST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20240626T180034Z
LOCATION:3010\, 3rd Floor
DTSTART;TZID=America/Los_Angeles:20240626T161500
DTEND;TZID=America/Los_Angeles:20240626T163000
UID:dac_DAC 2024_sess144_RESEARCH1233@linklings.com
SUMMARY:Arbitrary-size Multi-layer OARSMT RL Router Trained with Combinato
 rial Monte-Carlo Tree Search
DESCRIPTION:Research Manuscript\n\nLiang-Ting Chen, Hung-Ru Kuo, Yih-Lang 
 Li, and Mango C.-T. Chao (National Yang Ming Chiao Tung University)\n\nThi
 s paper presents a novel reinforcement-learning-trained router for buildin
 g a multi-layer obstacle-avoiding rectilinear Steiner minimum tree (OARSMT
 ). The router is trained by our proposed combinatorial Monte-Carlo tree se
 arch to select a proper set of Steiner points for OARSMT with only one inf
 erence. By using a Hanna-grid graph as the input and a 3D UNet as the netw
 ork architecture, the router can handle layouts with any dimensions and an
 y routing costs between grids. The experiments on both random cases and pu
 blic benchmarks demonstrate that our router can significantly outperform p
 revious algorithmic routers and other RL routers using Alpha-Go-like or PP
 O-based training.\n\nTopic: EDA\n\nKeyword: Physical Design and Verificati
 on\n\nSession Chairs: UDAY MALLAPPA (Intel Corporation) and Satish Sivaswa
 my (Advanced Micro Devices (AMD))
END:VEVENT
END:VCALENDAR