The problem of optimal assignment of autonomous vehicles in production-logistics system

Abstract

The problem of optimal assignment of autonomous vehicles to ongoing production processes is considered. Virtual testing for predictive maintenance involves the creation of digital twins based on simulation models. The solution of the assignment problem described in the article is the basis for building simulation models on Petri nets to analyze the dynamics of the production- logistics system. The structure of the agro-industrial system of grain harvesting by a complex of combines using autonomous vehicles based on KAMAZ vehicles is considered. Many characteristics of the production-logistics system have been determined: field areas and productivity, productivity of combines, carrying capacity and speed of vehicles, etc. The problem of minimizing operating costs for a complex of vehicles for a given distribution of combines over fields with grain crops has been formulated. The problem relates to integer linear programming with Boolean variables. The difference in the formulation of this problem lies in the formation of a number of restrictions that take into account the main parameters of the production-logistics system. An example of optimizing the distribution of autonomous vehicles for a given number of fields of harvesters processing them is considered. The resulting solution allows you to determine the set of vehicles ready for operation, as well as form a reserve of vehicles to reduce downtime. The solution of the proposed problem can be used as the basis for the structure of the simulation model as part of the digital twin of the production and logistics system. Also, the technique can be used in the current planning of real work of combines and vehicles.