The increase in the volume of information leads to an increase in the technologies and research for their processing. The study proposes a parallel implementation technology for matching records with the same characteristics of the algorithm, and the sequence of solving the problem using genetic algorithms is presented. In the proposed algorithm, the selection and replacement stages are determined by the parallel increase in the fitness function using replacement with sets. A genetic algorithm model is presented for the file design problem and the method of encoding chromosomes is explained. The process of working with existing methods for constructing the fitness function is described. At the end of the study, the reliability of the proposed model and algorithms is proven using experimental results