Digital predictor for business

Every day we use the services of many companies: we go to shops, cinemas, museums, we go to the subway, we fly by plane. Most of the companies we interact with on a daily basis are large ones. In fact, as a result of their activities, companies (private and state) create the world around us and create comfortable and safe conditions for consumers: delivery services strive to deliver goods even faster, stores are concerned about maintaining social distance.

The main difficulties that companies face when implementing such projects are the processes of servicing a large number of people and the uncertainty caused by the overlap of many factors.

Simulation modeling is one of the technologies used in such situations. Basically, it is a new approach to modeling for quickly creating complex models of complex business and social systems.

Unlike conventional modeling methods (e.g. Excel), in which we link cells to each other using formulas, the simulation model uses:

• agent-based modeling techniques (the model uses not numbers, but agents / objects that can move in space, interact with each other, exchange messages),
• discrete-event simulation (in the model, as in life, there is a time at which events related to each other occur),
• system dynamics (a method for describing the dynamics of complex business systems using streams and storages).

Simulation modeling allows you to play life scenarios when it seems that the computer is going through a process. Very complex systems can be simulated that are not easy to build into functions or predict because the number of participants can be very large. This improves the effectiveness of social security systems for both state-owned enterprises and businesses in the commercial sector.

Now there are many cases, both in Russia and abroad, when companies have managed to improve people's lives through simulation modeling.

For example, a project was implemented to model the Polytechnic Museum. The task was to simulate pedestrian flows in the museum, to understand where and in what quantity to place services: ticket offices, ticket terminals, wardrobes and even toilets in order to avoid crowds and queues, as well as increase the level of comfort for museum visitors.

Another hot topic is the return of people to the office after the pandemic. Many companies are interested in creating a model of office work in order to understand how to minimize the risks of infection of employees: how to properly organize the routes of people, minimize contacts.

The defense sector is very interested in simulating air attacks. Such models make it possible to form an optimal strategy for protecting the population in the event of an attack.

The urgent task of large FMCG companies is to reduce logistics costs. Simulation and optimization models allow you to find a chain configuration in which the level of customer service remains the same and the total logistics costs are lower. For example, now such a project is being implemented for a leading international beverage manufacturer.

A rather complicated project was the creation of a model of a workshop for the production of fasteners. The task was to increase production capacity. It was necessary to analyze investment risks and optimize new production processes.

Based on the collected data, an accurate digital model of the workshop was developed with an interactive interface for customizing scenarios. After 100 experiments with different parameter values, production gaps were identified. For example, while the model was running, the storage area in front of the electroplating line was 100% full. A year later, the space deficit exceeded 40%.

The uneven distribution of machines in the workshop and containers with blanks required a large number of loaders. The only girder crane in the workshop did not allow the production of products in the required volume due to breakdowns and repairs. To eliminate these problems, the filling of the workshop was changed in the model, a configuration was chosen in which the arrangement of equipment did not interfere with the release of the required volume of products. Also, recommendations were formulated for the installation of an additional girder crane and an increase in the number of loaders to increase the productivity of the entire workshop by at least 7%.

The two cornerstones of simulation modeling are:

• getting accurate inputs (the accuracy of the model depends entirely on this),
• choosing the right level of abstraction (model detail) appropriate for the task at hand.

For example, input data for modeling the flows of subway visitors can be obtained using IoT technologies (video analytics, various sensors, tags) or, in the old fashioned way, using on-site survey.

Building a digital twin based on simulation modeling enables businesses to predict and create competitive advantage.

Applying in practice the technology of simulation modeling, it is possible to calculate the behavior of an object or object at different stages of life, insure against danger and minimize risks.