Savchenko L. Increasing the prediction efficiency in transportation systems

Analytical relation of the correlation function and time interval function of transportation process has been established. Model for transportation process (the dynamic filter) which is effected by the "white noise" disturbance has been proposed. Moreover, the forming dynamic filter's order corresponds to the Erlang distribution's order which is modeling the real time interval function of transportation process. Forecasting model for unstationary transportation process used parametrical correlation function which applies to look through the unstationary transportation process during interval of prediction as theoretically stationary transportation process has been proposed. Experimental test devised models during the prediction of transport flow's characteristics in automated systems of traffic management and prediction of the stock remnant's value had proved the theoretical models were adequate. Key words: predictions, spectrum-correlation models, transportation process, systems of traffic management, planning of transportation.