作者: Peter Stone , Kurt Dresner
关键词: Real-time computing 、 Simulation 、 Traffic bottleneck 、 Reservation 、 Computer science 、 Metric (mathematics) 、 Traffic flow 、 Intersection (aeronautics) 、 Traffic congestion 、 Traffic congestion reconstruction with Kerner's three-phase theory
摘要: Traffic congestion is one of the leading causes lost productivity and decreased standard living in urban settings. Recent advances artificial intelligence suggest vehicle navigation by autonomous agents will be possible near future. In this paper, we propose a reservation-based system for alleviating traffic congestion, specifically at intersections, under assumption that cars are controlled agents. First, describe custom simulator have created to measure different delays associated with conducting through an intersection. Second, specify precise metric evaluating quality control Using metric, show our can perform two three hundred times better than lights. As result, it smoothly handle much heavier conditions. We very closely approximates overpass, which optimal solution problem dealing.