Search

This article presents a novel approach for the dynamic control of a signalized intersection.At the intersection, there is a number of arrival flows of cars, each havinga single queue (lane). The set of all flows is partitioned into disjoint combinationsof nonconflicting flows that will receive green together. The dynamic control of thetraffic lights is based on the numbers of cars waiting in the queues. The problemconcerning when to switch (and which combination to serve next) is modeled as aMarkovian decision process in discrete time. For large intersections (i.e., intersectionswith a large number of flows), the number of states becomes tremendouslylarge, prohibiting straightforward optimization using value iteration or policy iteration.Starting from an optimal (or nearly optimal) fixed-cycle strategy, a one-steppolicy improvement is proposed that is easy to compute and is shown to give a closeto optimal strategy for the dynamic problem.