Application of a subnetwork characterization methodology for dynamic traffic assignment

Abstract

The focus of this dissertation is a methodology to select an appropriate subnetwork from a large urban transportation network that experiences changes to a small fraction of the whole network. Subnetwork selection techniques are most effective when using a regional dynamic traffic assignment model. The level of detail included in the regional model relieves the user of manually coding subnetwork components because they can be extracted from the full model. This method will reduce the resources necessary for an agency to complete an analysis through time and cost savings. Dynamic traffic assignment also has the powerful capability of determining rerouting due to network changes. However, the major limitation of these new dynamic models is the computational demand of the algorithms, which inhibit use of full regional models for comparing multiple scenarios. By examining a smaller window of the network, where impacts are expected to occur, the burden of computer power and time can be overcome. These methods will contribute to the accuracy of dynamic transportation systems analysis, increase the tractability of these advanced traffic models, and help implement new modeling techniques previously limited by network size. The following describes how to best understand the effects of reducing a network to a subarea and how this technique may be implemented in practice.