Dynamic and agent-based models of intelligent transportation systems

The authors present mathematical and simulation models of intelligent transportation systems (ITS). The models of two types are considered: the dynamic model of cargo transportation and agent-based model of the ITS - the 'Manhattan grid' type. The problem of rational railway planning related to research of cargo transportation models and corresponding cargo flows within the dynamic system is studied. The process of cargo transportation was modelled considering the mechanism of interactions with major railway infrastructure elements. The variation ranges of parameters at which cargo transportation system can be consistently active are defined. Possibilities of simulation modelling transportation and pedestrian flows at the micro-level considering complex interactions between heterogeneous agents, in particular, vehicles-to-pedestrians (V2P), vehicles-to-vehicles (V2V), vehiclesto-infrastructure elements (traffic lights) (V2I) etc. using the case study as the ITS belonging to the "Manhattan grid" type studied. As a result, it is shown that ITS with partially controlled pedestrian crossings have advantage by the level of the total traffic in comparison to the ITS with uncontrolled crossings, especially with low-intensity and high-speed traffic. The two types of models are united by the unity of their tool-making description. For models of the first type, all processes at the micro-level are strictly regulated. Therefore, such systems are well characterized by established macro-indicators - states of the soliton solutions class (i. e. the solutions of travelling wave type). In models of the second type, there are large fluctuations at the micro-level that affect the safety of road users (e. g., traffic jams, accidents, etc.). This explains the use of agent-based models that consider processes at the micro-level. At the same time, macro-indicators are the most important characteristics for checking the adequacy of agent-based models.