Integrated assessment of traffic management efficiency in real time based on DTA model

Abstract

Currently, traffic engineering and management methods use many traffic management tools in the form of direct or indirect management systems. In general, a steady transition to intelligent traffic management systems is observed. Any intelligent system is comprised of a large number of subsystems, including a variety of computer dynamic real-time models. The simulation methods used in intelligent transportation systems (ITS) basically can be divided into three classes: dynamic traffic assignment (DTA) models, machine learning-based models with a set of different statistical submodels, models using a hybrid approach including a series of meso-DTA models, micro-level models, and machine learning. The main task of the above-mentioned models is to provide high-quality short-term traffic forecasts with a small estimated time. Such models make it possible to choose the most effective strategies of traffic management based on a variety of data obtained from peripheral devices and control actions with the available tools comprising the ITS. The application of such models requires a constant assessment of the efficiency of particular management scenarios. Usually, it is required to evaluate 3–5 scenarios of traffic management within a short period. Thus, the development of a method for the integrated assessment of traffic management efficiency when simulating traffic management scenarios as part of the ITS based on real-time models represents an important scientific challenge.