On March 23, a team from the Software Department of UNN Institute of IT, Mathematics and Mechanics won the international online Generalization-based Contest in Global Optimization GENOPT 2016. The team consists of: Assoc. Prof. Konstantin Barkalov, Senior Lecturer Alexander Sysoyev, Assistant Ilya Lebedev and student Vladislav Sovrasov.

Contestants in the online competition were invited to solve complex global optimization problems. GENOPT runs as a competition in three categories for students and researchers. Our team won in one category (GENOPT Target Shooting 2016) and came third in two other categories. Winners will present their algorithms during a special session in the LION10 conference, which will be held from May 29 to June 1, 2016, in Ischia Island, Italy.

Assoc. Prof. Konstantin Barkalov and Assistant Ilya Lebedev are involved in a research project on global optimization, HPC and supercomputing applications. The project, led by Prof. Yaroslav D. Sergeyev, is funded by the Russian Science Foundation. Theoretical and experimental studies are carried out to develop new integrated models and breakthrough methods for global optimization problems as well as applications that run at trans-petaflop speeds.

Global optimization is a rapidly developing science that is aimed at finding globally optimal solutions to multiextremal optimization problems in the fields of economics, ecology, engineering design, identification science, image recognition, signal processing, navigation etc.

Development of effective methods of multiextremal global optimization is a task of high complexity, because it demands that researchers build and use integrated adaptive models for quality criteria dynamics based on test results that show the accuracy of application of these methods. Developers devise original schemes to reduce the dimension and complexity of the problems, use these schemes and models for effective planning of deployment of new tests. This also involves appropriate storage of information and application of nontrivial algorithms for parallel processing and distributed computing.

Find out more about GENOPT