Genetic algorithms with group sampling and directed mutation

摘要

Problems of optimization of complex multiextremal functions and corresponding methods of solu- tion are discussed. A group variant of genetic algorithms is considered and a new type of mutation that allows one to make an algorithm more specific and effective is proposed. A detailed description of the algorithm and a class of problems for which this algorithm is most efficient are presented. A classification of types of problems for the choice of an appropriate algorithm for their solution is carried out with the use of experimental results.