BRYANSK, October 31. /TASS/. Researchers of the Bryansk State Technological University of Engineering have developed a system of business processes forecasting based on neural networks, Pro-rector of the University Oleg Kazakov told TASS.

"A technology of building up digital twins of business processes and a mechanism of managing real business operations with their aid were developed. This was implemented as part of the so-called Industry 4.0 concept, which implies the large-scale rollout of information technologies in the industry and automation of business processes," Kazakov said.

The project received the grant support from the Russian Scientific Foundation amounting to 1.2 mln rubles ($13,000) and the technology was tested at the Bryansk Automobile Plant. The smart system of business processes forecasting will make it possible to solve in a new way key production tasks of industrial plants with the use of digital twins. "The new method of automatic functional modeling of enterprise processes makes it possible to automatically build a functional model of a business process through integrated use of author algorithms based on graph neural networks and procedures of converting the business process model extracted from raw data," the developer noted.

The studied technology enables implementing digital twins of processes in different areas, from control of process operations and logistics to management accounting. "For example, as part of inventory management at an enterprise, a reliable virtual copy of this process is created automatically. It is presented visually to a user on a monitor as a multilayered 3D structure, where each layer visualizes a certain control aspect: the sequence of performing transactions, engaged resources, and so on. A digital twin operator can in real time execute a command of launching the procedure of inventories moving from a storage area to production shop," he added.