ОПТИМИЗАЦИЯ УПРАВЛЕНИЯ ГОРОДСКИМИ СВЕТОФОРАМИ С ПРИМЕНЕНИЕМ НЕЙРОННЫХ СЕТЕЙ

Abstract

One of the most pressing problems of large cities is the problem of traffic management of vehicles. The reason for this problem is an imperfect way to manage traffic flows. Traffic light regulation is of particular importance in traffic management. Most modern traffic light control systems operate at set time intervals and are not able to cope with the constantly changing situation on the road. A promising direction for solving this problem is to optimize the system using artificial neural networks. The advantage of neural networks is self-learning, which allows the system to adapt to the changing situation on the road. Despite numerous attempts, it has not yet been possible to obtain a high-quality mathematical model of urban traffic management. This model should determine the functional dependence of transport flow parameters on control parameters. Nowadays, traffic flows are regulated everywhere by means of traffic lights. If we can get a fairly accurate mathematical model of traffic flows, we can determine the optimal duration of the traffic signal phases to achieve the maximum capacity of the road network node. A fairly accurate mathematical model of traffic management that works in predictive mode will display an estimate of the optimal control parameters, as well as make correct decisions in emergency situations. Well-known mathematical models of road traffic take into account only the average values of traffic flows, and not the exact number of cars on each road section at a particular time.