Taking into Account the Variable Nature of the Calculation Standards of the Train Formation Plan Using Computational Intelligence Technologies

Abstract

Purpose: The train formation plan is the most important logistics tool for cargo transportation management. But the use of discrete values of design standards in its development — the quantity of car-hours for the accumulation of trains and savings from passing of cars through technical station without processing — do not always guarantee the optimal solution due to the objectively existing unevenness of operational work. It is precisely this factor that generates uncertainty, which is not inherently stochastic, and necessitates adjustments to the train formation plan throughout its lifecycle. In addition to fluctuations in car traffic, it causes a change in the values of the calculated standards. To manage such uncertainty, there are special methods, one of which is fuzzy logic. The article describes a way to account for changes in all calculation standards of the train formation plan. Methods: The methods of one of the technologies of computational artificial intelligence are used — fuzzy logic, fuzzy sets and fuzzy mathematics. Results: The dependence has been established on how changes in the calculated parameters of the train formation plan are affected by fluctuations in car flows on specific destinations. The formulas obtained on the basis of known analytical expressions allow us to determine the values of the standards of the formation plan without using auxiliary tables and graphs. Practical significance: The use of the obtained dependencies in the development of the formation plan will improve the accuracy of its calculation by taking into account fluctuations not only of car traffic, but also the values of the calculated standards depending on them.