{"id":"arxiv/2610.06266","url":"https://arxiv.org/abs/2610.06266","tags":["cs.LG","cs.AI","math.OC"],"score":0,"title":"Ramp Metering Control via Hybrid State Deep Reinforcement Learning in Partially Observable Connected Vehicle Environments","domain":"cs.LG","source":"arxiv","authors":["Youcef Mehamlia","Nadir Farhi","Meriem Bouali"],"reviews":0,"subject":"cs.LG","summary":"Freeway on-ramp merges are major sources of congestion, causing significant economic and environmental costs. While Deep Reinforcement Learning (DRL) offers a promising solution for ramp metering, existing approaches rely primarily on aggregated macroscopic data. Connected vehicles (CVs) provide vehicle-level observations that can complement aggregate traffic measurements, but their limited penetration produces incomplete microscopic information. This paper proposes a hybrid observation representation combining macroscopic traffic measurements with a two-channel grid encoding observed CV…","category":"research","comments":0,"disputed":false,"division":"cs","mentions":0,"discussed":0,"published":"2026-10-05","reactions":{}}