Планирование оптимальной загрузки железнодорожного транспорта горнодобывающей компании

Abstract

The article considers the problem of planning the optimal loading of freight trains on a limited railway network – the network of a mining company. Two variants are illustrated. In the first variant only one type of locomotives operates on the railway network. In the second variant there are different types of locomotives. The types of lo-comotives are determined by their carrying capacity. The algorithm of scheduling multiple locomotives in a network is a combination of solutions for particular scheduling problems for each locomotive subject to the restrictions on their joint movement. Structurally, the set of possible locomotive schedules is a set of system states at discrete points in time. The state at each moment of time describes the location of each locomotive at a particular station and the amount of cargo transported. Thus, two consecutive states unambiguously determine the routes of locomotives. To describe a complete picture of all possible states and the sequence of their transitions from one state to another are represented in the form of a tree. There is proposed an algorithm for decreasing the variants of schedules by comparing the states in the tree nodes at each step and excluding the branches of states that would knowingly lead to a non-optimal schedule. The problem statement including several locomotives of different types, the solution algorithm and its application to a particular case are presented. In the formulation of the task the criterion of optimality for comparing different schedules is the minimum completion time of transporting the necessary volume of cargo. The additional criterion is total operating time of all locomotives, which is used only in case of equality of the main criterion for several schedules.