Author(s)

Parham Tavakolian

Advisor(s)

Peter Gregory Furth

Contributor(s)

Daniel M. Dulaski

Date of Award

2011

Date Accepted

2011

Degree Grantor

Northeastern University

Degree Level

M.S.

Degree Name

Master of Science

Department or Academic Unit

College of Engineering. Department of Civil and Environmental Engineering.

Keywords

Optimal Sequencing, Traffic Engineering

Subject Categories

Signalized intersections, Traffic engineering

Disciplines

Civil Engineering

Abstract

At a signalized intersection, efficiency can be gained by running compatible traffic streams in parallel, while incompatible streams run in series. In an intersection with an asymmetric clearance times, some sequences require less clearance time than others. In a simple intersection with a limited number of sequences, finding the optimal sequence is trivial. However in a complex intersection with a large number of traffic streams or unusual geometry, the optimal sequence is not obvious, and the number of possible sequences can be too large to check manually. An optimal sequence increases intersection capacity and minimizes delay by allowing an intersection to run with a shorter cycle length while serving the demand of all the traffic streams

Previous research has proposed methods for finding the optimal sequence. However, counterexamples given in the literature as well as in this thesis show that their algorithms sometimes fail to find the optimal sequence. This thesis applies exhaustive implicit enumeration using the branch and bound algorithm together with a network linear programming (LP) formulation for finding the minimum cycle length for a given sequence. Network LP is extremely efficient. Branch and bound is efficient because it uses the concept of conflict groups to limit the search to feasible sequences. In addition, the algorithm supports simultaneous start and staggered start constraints, along with the usual conflict constraints.

For Chapel Hill, a test intersection used in the literature, our method finds not only the same solution reported by others, it finds a better solution, with a 12.5% reduction in necessary cycle length and 18% increase in capacity. In a test of a complex U.S. intersection, Charles Circle with 18 traffic streams, the algorithm found a more efficient solution than those known previously. Computer processing time for all examples was under 3 seconds.

Document Type

Master's Thesis

Rights Holder

Parham Tavakolian


View Full Text

Click button above to open, or right-click to save.

Share

COinS