| Summary: | Background. An approach to the formalized representation of models of genetic
algorithms as a kind of evolutionary calculations is proposed. The issues of data representation to improve the efficiency of evolutionary search, the choice of operations in the formalized
representation of genetic transformations and the construction of mathematical
models of calculations are discussed. It is noted that the organization of evolutionary calculations
within the framework of the most common varieties of genetic algorithms leads to
the need for a formalized representation of a sufficiently large set of parameters that affect
the efficiency of the process of evolutionary calculations in relation to a specific task. Two
approaches to the task of binary representations are considered. The first approach assumes
a binary representation, the second approach assumes a binary encoded representation.
Models for unlimited and restricted populations providing the use of a wide range of transformation
and selection operators are discussed. The proposed approach is most effective in
solving the problem of evolutionary calculations, taking into account the maximization of
key patterns. Results. The theoretical substantiation of the structure and parameters of models
of evolutionary calculations for solving problems of identification and forecasting of the
state of technical objects and socio-economic systems is carried out. Conclusions. The proposed
approach provides a formalized description of the basic transformations for models
of evolutionary calculations with limited and unlimited populations. The approach can be
used in the formalized formulation of the problem of maximizing key patterns in an evolving
model.
|
|---|