UC Berkeley

Personal rapid transit (PRT) is a concept in which a fleet of automated vehicles operate on a network of grade-separated guideways, accessible on-demand at a fixed number of stations. The questions this research addresses include: How can merge-induced congestion be modeled and controlled? How do station configurations influence delay? How large must fleets be to accommodate a given demand profile? How should empty vehicles be dispatched to rebalance inventories at stations? While current literature on PRT tends to rely on simulation, this thesis uses queueing theory as an analytical framework.