Digital Design creates the solution for automation of logistic engineering in Gazprom Neft Scientific Technical Center

On December 3, 2018 Digital Design reported that it was selected by the contractor of carrying out Research and Development of Gazprom Neft Scientific Technical Center LLC for the purpose of further development of a prototype of the solution "Integrator of Conceptual and Logistic Engineering".

Service which will allow to consider transportation, warehousing and other transactions made on delivery the material resources (MR) at a stage of conceptual design of arrangement of the field should become a project deliverable. In other words, service will help to calculate how quicker and cheaper to deliver different construction materials to fields.

The cost of logistics will be calculated on the basis of a set of factors: data on regional cargo flows, their capacity, transport communications, geographic location of points of storage and others. A system should build automatically a matrix of options taking into account all data on three cost groups: transport, capital and operational.

Important part of the solution of a problem of creation of optimal plans of capital construction is determination of the logistics chains including both transport, and warehouse logistic tasks. And it is necessary to solve these problems so that normative delivery dates at the minimum total cost were observed. It is possible to solve it by method of "brute force", touching all possible options, but at a large number of nodes of network such search can take a lot of time therefore it is not suitable for practical use. There are different approaches to the solution of a logistic task which are based on selection of the certain heuristics imitating actions of the specialist. But not always this approach yields optimal result because the specialist usually solves a problem based on the accumulated experience. In scientific laboratory algorithms based on machine learning, in particular "training with a reinforcement" are developed. This class of algorithms allows to create the self-training systems which are used during creation of weak artificial intelligence for creation of a management system for unmanned vehicles, chat-bots, bots for management of characters in computer games. The same approach can be used and for solving of tasks of logistics when we train an algorithm to apply the optimal and right decisions to minimization of target function of the total cost of logistics at design of the field. Ilya Ashikhmin, head of scientific laboratory "Digital Design"