This dissertation furthers the understanding of macroscopic traffic behavior on urban networks. In particular, it focuses on the aggregate effect of turning maneuvers in networks with multiple routes. The presence of turning between different routes on a network is found to have two effects: 1) it causes chaotic and inefficient behavior, especially when the network is congested; and, 2) it may reduce maximum vehicle flows across the network. Fortunately, this work finds that designing networks that are redundant (i.e., have multiple unique routes between all origin-destination pairs), helping drivers avoid locally congested regions and limiting the rate at which vehicles are allowed to enter the network can help mitigate the first effect, and this allows the network to operate more efficiently. It is also found that the second effect may not always be harmful---lower network flows do not necessarily result in decreased network efficiency if the lower flows are accompanied by more direct routing between origins and destinations. In fact, two-way networks, which accommodate conflicting left-turns and result in lower maximum vehicle flows than one-way networks, are found to serve trips at a higher rate because drivers travel shorter distances on average. Thus, in many cities, maximum network efficiency can be improved by converting one-way streets to two-way operation.
The different versions of the original document can be found in:
Are you one of the authors of this document?