Analysis of one-way and two-way street configurations on urban grid networks

Abstract

The design of urban street networks is a subject of much controversy because there is no consensus on the best way to organize streets to maximize mobility. Debates between one-way and two-way street networks have existed since cars were introduced into urban environments and are still ongoing. In this paper, we provide additional insights on this issue by analyzing traffic behavior on three different street configurations: one-way streets, two-way streets, and two-way streets with prohibited left turns. The outcomes of this paper will furnish urban planners and policy makers with a better understanding of the mobility provided by each of these types of urban street network layouts. <br/><br/>This study is focused on abstract finite grids operating in both uncongested and congested conditions. Analytical formulations are created and used to compare the three network types under low demand scenarios when the network is uncongested. These formulations rely on a fixed traffic assignment scheme in which drivers are assumed to minimize their overall travel distance and number of turning maneuvers. They are used to reveal general insights. In addition, simulation techniques are employed with a static traffic assignment algorithm to examine the behavior of networks under more realistic congested situations. <br/><br/>In general, the study reveals that two-way networks provide the shortest distance traveled and result in more uniform congestion patterns due to route redundancy. However, they are severely penalized by restrictions in capacity at the intersections due to more complicated phasing schemes, which leads to higher average travel times when the network is large, has short links, or is lightly congested. One- way networks offer the longest travel distances but have intersections with very high capacity. They perform better when the network is very large or links are very short. Finally, two-way networks with prohibited left turns provide the best compromise between distance traveled and intersection capacity, which makes them suitable for many different scenarios with low and medium congestion. Their main weakness is that they suffer from inhomogeneous congestion patterns caused by a lack of route redundancy. This could lead to very large travel times when the network is very congested.