We use cookies.
By using the site, you agree to our Privacy Policy.

Laboratory «Hybrid modeling and optimization methods in complex systems»

Contract number
Time span of the project
General information
Name of the project:

Hybrid methods of modeling and optimization in complex systems

Goals and objectives
Goals of project:

Creating a laboratory that will develop and apply metaheuristic and other hybrid methods of general-purpose optimization.

Project objective: 

  1. Creating a technology for automated algorithmic design that will ensure an efficient combination of precise and mathematically substantiated methods methods with heuristic methods of solving complex optimization problems.
  2. Creating an education center on the basis of the laboratory to implement education programs as well as separate modules on the mathematical modeling and optimization of complex systems, including with application to machine learning and artificial intelligence systems.

Research directions: Computer, information sciences and technologies

The practical value of the study
Planned project results:

  1. Methods and algorithms of solving problems of optimization and modeling that combine heuristic/metaheuristic approaches to optimization and modeling problems with precise and mathematically substantiated optimization methods.
  2. Approaches to the adaptive decomposition of problems of ultra-high dimensionality and hybrid methods of solving those problems based on self-adaptive differential evolution algorithms and methods of local search, algorithms of identification and analysis of separable and intersecting components of ultra-high-dimensionality problems .
  3. Parallel algorithms for ultra-high-dimensionality problems with the use of competitive and cooperative coevolution, insular models and multipopulation models based on computational clusters (grid systems and GPU clusters) and results of analysis of their comparative efficiency.
  4. New optimization methods based on the hybridization of subgradient minimization methods and machine learning algorithms with gradient descent algorithms.
  5. Algorithms of differential evolution and general-purpose computations based on the use of graphics processing units for numerically efficient and symbolic implementation of dynamic systems, aggregation with optimization recurrent neural networks.
  6. Fuzzy logic systems for the development of efficient models for control by solving nonstationary and stationary problems of nonlinear optimization.
  7. w3Hybrid algorithms for the optimization and logical analysis of data for solving problems of the classifications of objects in interpretable machine learning.
  8. A technology for the complex modeling of compound systems based on creating a space of models with the use of self-adaptive hybrid methods of optimization and expert knowledge as well as methodologies for searching through this space that allows to produce ensembles of models of compound systems and processes.

Hide Show full
Other laboratories and scientists
Hosting organization
Field of studies
Invited researcher
Time span of the project
Research of network technologies with ultra-low latency and ultra-high density based on extensive use of artificial intelligence for 6G networks

The Bonch-Bruevich Saint Petersburg State University of Telecommunications

Computer and information sciences

St. Petersburg

Abd El-Latif Ahmed Abdelrahim



Laboratory for Non-linear and Microwave Photonics

Ulyanovsk State University - (USU)

Computer and information sciences


Taylor James Roy

United Kingdom, Ireland


Multi-scale Neurodynamics for Smart Systems

Skolkovo Institute of Science and Technology - (Skoltech)

Computer and information sciences


Wang Jun