A path-based analytical dynamic traffic assignment model for real time evaluation of simulation based dynamic traffic assignment models

Abstract

In the development of a deployable real-time DTA system, macroscopic analytical DTA models are employed to provide credible benchmarks for evaluating simulation-based DTA models. To achieve this goal, an analytical link-based DTA model and a path-based DTA model are developed to incorporat the rolling horizon implementation, traffic control models, and on-line calibration process. A benchmark evaluation for simulation-based DTA models is the intended agenda of the proposed model, algorithm, and implementation approach. Specifically, a path-based DTA model is discussed in this paper to illustrate the complicated process of model and algorithm development. This model is formulated as a variational inequality (VI) and can be solved efficiently to convergence by the proposed relaxation algorithm. The incorporation of rolling horizon implementation, traffic control models, and on-line calibration makes the proposed model and algorithm more appropriate for bench-mark evaluation for simulation-based DTA models. According to different assumptions of travelers’ route choice behavior, travelers are classified into three different classes, including fixed (predetermined) route class stochastic dynamic user-optimal (SDUO) class, and dynamic user-optimal (DUO) class. The solution steps and the combined algorithm are discussed in detail. Computational results are also reported and analyzed.