Abstract: This article is devoted to econometric analysis of the results of experiments conducted with two agent-based models, which describe the movement of ground vehicles. There are two types of road users in these models: manned ground vehicles (MGV) and unmanned ground vehicles (UGV). In the first model, the main difference between UGV and MGV is an ability to exchange massages between UGV for transmitting information about extreme situations, which allows them to adjust speed and direction of movement. In the second model, in addition to the above differences, UGV have an additional advantage, namely, the ability to intelligently assess density of traffic flow for efficient maneuvering. In these models, at a given roundabout, traffic characteristics such as output stream traffic and the number of traffic accidents are analyzed. The main task of the econometric analysis is to study dependence of these traffic characteristics on the model parameters such as average vehicle speed, input flow rate, message exchange rate between UGV, and the impact of the effect obtained from the implementation into UGV ability of intelligent estimation of traffic flow density.
A new approach to the transformation of solutions of optimal control problems based on the special form of relaxation of complementary slackness conditions is presented. The proposed approach is tested on the Russian banking system model, which is derived as a solution of a linear nonautonomous optimization problem with mixed constraints. It is shown that the use of this method regularizes the model in a sense it becomes applicable for the forecasting of the main Russian banking indicators.