With hardware getting cheaper on daily basis, high-performance computing has become one of the more compelling options for running advanced application programs efficiently, reliably and quickly. The research is oriented towards usege of many-core systems, e.g., modern multiprocessors and general-purpose GPUs.
Complex systems modeling and black-box optimization
Modeling is a conceptual tool widely used in science and engineering to formalize the abstract notions of reality. It provides a framework that can be evaluated and reused for reasoning in a wide range of situations. In fact, models can be seen as simplified representation of complex reality, that can help us understand, analyze, control or predict the behaviour of the considered real processes and phenomena. Such models are suitable for further processing in form of black-box optimization.
Metaheuristic optimization techniques are nowadays being researched and employed in practice at an increasing rate. Their robustness, ability of providing multiple solutions to a problem at hand, and suitability for implementation in parallel computing systems make them both a challenging research issue and a powerful problem solving tool.
Bioinspired optimization algorithms are important for solving hard combinatorial and numerical problems in various domains of theoretical interest and practical applications. We study, develop, and deeply research different variations of naturally inspired algorithms and their ability to search for a solution with a promising and effective heuristic approach: genetic algorithms, ant-colony optimization, particle swarm optimization, neural networks, etc. We use these approaches for solving various multi-objective and constrained optimization problems from industry and real life.