Validation of an optimization model based stochastic traffic flow fundamental diagram

Abstract

The fundamental diagram is used to represent the graphical layout and determine the mathematical relationships among traffic flow, speed and density. Based on the observational speed-density database, the distribution of speed is scattered in any given traffic state. In order to address the stochasticity of traffic flow, a new calibration approach has been proposed to generate stochastic traffic flow fundamental diagrams. With this proposed stochastic fundamental diagram, the residual and stochasticity of the performance of calibrated fundamental diagrams can be evaluated. As previous work only shows the validation of one model, in this paper, we will use field data to validate other stochastic models. Greenshields model, Greenberg model, and Newell model are chosen to evaluate the performance of the proposed stochastic model. Results show that the proposed methodology fits field data well.