Engineering of the Future Prize bestowed on St Petersburg University Professor Oleg Granichin

We asked Professor Granichin to tell us about the collaboration between scholars at St Petersburg University and the Scientific and Educational Centre in the development of multi-agent technologies and their practical application.

Professor Granichin, could you please tell us what multi-agent technologies are in essence?

Multi-agent technologies are a kind of alternative take on what is known as Big Data and complex problems. Big Data and complex problems become so when you look at them as a whole. However, we divide them up into their components, into a set of local tasks.

Actually, the human brain and living colonies of organisms — fish, birds and ants — are constantly dealing with issues of Big Data. Take, for example, a flock of birds. On the one hand, it is a huge database, which includes the process of flying, the flight path, and the speed. On the other hand, each bird flies on its own. It’s the same with a community of people: it would seem that there are a lot of us, but we each go our separate way. A bird in a flock, a person in a crowd — we distinguish a dynamic objective in a big task, the actor of which can go it alone. The same sort of thing happens in multi-agent technologies.

How long has this avenue of research been developing at St Petersburg University?

Our research dates back to 2009. That’s when the term "multi-agent technologies" began to be used at the University, and I was probably the first to write about them in the way that they are understood today.

The Scientific and Educational Centre "Engineering of the Future" is a community of educational and research organisations along with organisations in the real sector of the economy. Its goal is to achieve world leadership in research and technological development and the training of personnel in the areas of activity that the centre is involved in. In December 2020, Engineering of the Future became one of the five such centres to receive world-class status.

This area of research is developing rapidly at St Petersburg University at the Mathematical Robotics and Artificial Intelligence Research and Educational Centre. This is a laboratory composed mainly of undergraduate and postgraduate students from the University. In our research on the application of multi-agent technologies, we concentrate on solving robotic problems. Today, this laboratory is recognised at a high level: at the most prominent Russian conference in the field, the Multi-Conference on Management Problems in Divnomorskoe Selo, Gelendzhik, we were acclaimed for presenting the best paper on the role of robots.

Most of the projects that we carry out are supported by grants from the Russian Science Foundation and the Russian Foundation for Basic Research. So, the Scientific and Educational Centre is a unique testing ground that engages young people in research and in solving interesting problems in practice.

How did the collaboration develop between the University and the Scientific and Educational Centre?

There have been strong ties between the research team at St Petersburg University and Samara University for a long time. Five years ago, we enrolled students in a joint target programme, and that’s when we combined forces and began working on multi-agent technologies. Together we developed terminology, approaches and concepts, and in 2021 Samara won a tender to set up a world-class scientific and educational centre.

After that, we continued to work together, but now within the framework of the Scientific and Educational Centre "Engineering of the Future", which encompasses several different areas of research. One of the fastest growing and strongest among them is Artificial Intelligence. It is in this area that the researchers at St Petersburg University have concentrated their efforts.

For what contribution to the development of multi-agent technologies were you awarded the Engineering of the Future Prize?

In recent years, we’ve come up with a whole theory about which Big Data problems can be divided into small, local ones for the agents. The most important thing is when the solution of these agents leads to a correct solution to a common problem. It turns out that if you limit the teamwork of the actors, for example if you come up with a model in which each bird in a flock is connected only to its neighbours, and in a crowd of people each of them deals only with the two who are closest, the computational complexity is not all that great.

Such a solution is not suitable for all problems, but there is a huge class of them that are connected with a conflict of interests and with choosing the optimal strategies, for which it is. Very often, when there are common challenges, the approach of the well-known Nobel laureate John Forbes Nash Jr is used. According to his theory of Consensus data, when you have disparate interests, the optimal solution is the one that is optimal for all. No one can improve their strategy at the expense of others, so in this way it is good for everyone. It turns out that to solve such a class of problems, multi-agent technologies are more than sufficient. If a problem comes down to a question of consensus, such a simplification and local interaction are enough to arrive at a common optimal solution in a simple way. This is our contribution to the formation and development of multi-agent technologies.

What problems do St Petersburg University researchers resolve in their work for the Scientific and Educational Centre?

With the help of the latest technologies, we launched the unmanned cargo spacecraft Progress. Multi-agent technologies are mainly used as software systems, and also to solve scheduling problems, for example to load the International Space Station from the Progress spacecraft. The loading and the calculations are carried out using new computational technologies.

Among the co-founders of the World-Class Scientific and Educational Centre "Engineering of the Future" are Samara State University, St Petersburg University, Samara Technical University, and Peter the Great St Petersburg Polytechnic University. The centre’s industrial partners include: The State Corporation for Assistance to Development, Production and Export of Advanced Technology Industrial Product ‘Rostec’; The State Space Corporation ‘Roscosmos’; Russian Railways; and other organisations.

Multi-agent technologies are also used to develop railway timetables. And, under the aegis of the Scientific and Educational Centre in Samara, there is a project in the works to create a constellation of artificial satellites in the foreseeable future. Once launched, these satellites will interact with each other.

The Scientific and Educational Centre project is a successful synthesis of St Petersburg University scholars with their colleagues at other universities and organisations that are a part of the Centre. It gives us a chance to not only develop theories but also to put the results of our research into practice. Our researchers dig into the theoretical fundamentals, while our colleagues in Samara do the bulk of the programming and the engineering and technological work.

How do you assess the significance of this prize for St Petersburg University?

It means that we have chosen good partners, ones who trust us and give us an opportunity to move forward. It also underscores the authoritative reputation of our University.