Building a model for planning rapid delivery of containers by rail under the conditions of intermodal transportation based on robust optimization

This paper considers the possibility of devising a technology of fast railroad communication for the transportation of containers between the port and customer enterprises in the course of intermodal transportation. The purpose of technology development is to reduce the share of the use of trucks on intermodal routes and thus solve a number of related environmental, transport, municipal, and economic problems. The devised technology is based on the principles of bringing the railroad as close as possible to the end points of the route, minimizing the number of intermediate modes of transport, and enabling the maximum speed of movement of containers by rail. For this purpose, the use of MetroCargo™ freight terminals and CargoSprinter modular trains is proposed. In the course of the study, the task to reliably plan the operation of the fleet of such trains for the delivery of containers between the port and enterprises under the conditions of "noisy" initial data was set and solved. To this end, the problem was formalized in the form of a model of mixed programming, based on the principles of robust optimization. To optimize the model taking into consideration the principles of robustness, a procedure was proposed that uses a two-circuit genetic algorithm. As a result of the simulation, it was found that the resulting plan was only 6.5 % inferior to the objective criterion of the plan, which was compiled without taking into consideration robustness. It was proved that the devised model makes it possible to build an operational plan for the delivery of containers by rail, which is close to optimal. At the same time, the plan is implemented even in the case of the most unfavorable set of circumstances in the form of delays, shifts in the time windows of the cargo fronts, etc., that is, under the actual conditions of the transport proces