Multi-scale Neurodynamics for Smart Systems

Name of the project: Multi-scale intelligent neurodynamic systems for multi-dimensional optimisation in the field of machine learning and data processing

The scientific goal of the proposed project is the development of fundamentally new optimisation algorithms based on multi-scale collaborative neurodynamic systems and modern scientific approaches related to them. A special attention will be paid to the development of methods of solution of the following complex optimisation problems:

• Problems involving non-convex and discontinuous target functions and special limitations that guarantee smoothness and sparsity of the solutions;

• Problems involving integer and continuous variables produced from constrains on the allowable set of solutions;

• Problems of multi-criteria optimisation with multiple target functions in which it is necessary to efficiently find the solution suitable for all the target functions;

• Problems of multi-level optimisation in a hierarchical decision-making structure as well as multi-period optimisation with dynamic modelling in the context of a decreasing time horizon of planning that will be investigated from the point of view of dynamic optimisation.