Improvement of the method of dynamic routing with time windows of motor tracking of agricultural products

Abstract

The article is devoted to improvement of the method of dynamic routing of motor vehicles. Peculiarities of transportation of agricultural goods are considered. There is a high probability of improper transportation of perishable goods up to refusals to provide such services and loss of cargo and damage. More specifically, the problem of vehicle routing concerns the design of multiple road routes, the implementation of which is characterized by minimum costs incurred in meeting the demand for agricultural products of a group of geographically dispersed consumers, satisfying a group of operational constraints.

An overview of known publications related to the dynamic routing methods has been made. Three main factors of the need for dynamic tasks are taken into account. As the impact of these factors becomes stronger, development of online transport management methods is becoming more relevant. In addition, the analysis of known methods of routing and scheduling of vehicles showed that with the increase in the number of orders, especially unplanned, that methods became inefficient in the quality of results. Application of pre-classification is a way to improve the result.

The problem of dynamic routing is formulated as follows. There are many known orders for transportation of perishable goods. Orders must be completed as soon as possible. Delivery of these goods from agricultural producers to procurement enterprises is of small scale. The volume and type of cargo and starting or ending points of routes for each of the orders are set. Time windows of orders are formed under the influence of acceptable terms of harvesting and processing of agricultural products, on one hand, and consumer demand, on the other hand. The limits of the time window are the earliest point in time before which the order cannot be executed and the latest point in time after which the order must be executed. The content of the order is that the transport company must load the cargoes into the trucks within the time specified by the agricultural producer and deliver them to customers just in time. The main criterion for the quality of transportation is the minimum guaranteed duration of delivery of goods from the entire set of orders for transportation. Restrictions are applied to the forecast horizon, as well as time windows for cargo delivery. Another limitation is the intensity of the vehicle fleet use. The requirement of the minimum number of involved trucks from available is set. The proposed method of compiling dynamic schedules and routes differs from the known ones. The sets of input data in this method (transport orders) are first pre-classified according to the compatibility of processing in a single transport process. Then the schedule and route can be made using linear programming methods that provide a guaranteed accurate solution. In addition, the results of optimization of routes and timetables of trucks by the proposed method are stable and resistant to changes in transport and road conditions. Tests of the method on the test model showed that the method of branches and boundaries with a preliminary classification shows approximately 11 % less guaranteed delivery time of all goods. At the same time, the use of trucks also increases by 11 % and their unproductive mileage is reduced.