Effective algorithms for the synthesis optimization of a set of irredundant diagnostic tests on the basis of genetic algorithms in the intelligent system

In the paper effective algorithms for the synthesis optimization of a set of irredundant diagnostic tests with genetic algorithms used to solve problems of a large dimension are suggested. Effective algorithms are carried out in two stages. The initial stage involves the synthesis of perspective populations on the basis of creating the irredundant partial implication matrix sectionalized by classification mechanisms; revealing certain kinds of regularities, which are used combined with genetic transformations for creating a set of irredundant diagnostic tests. In the second stage optimization of a set of already constructed irredundant diagnostic tests is carried out on the basis of selecting pseudoobligatory genes (features) from a set of perspective chromosomes (a set of irredundant diagnostic tests) constructed at the first stage. The second stage is performed by one of two step-by-step algorithms. All the obligatory, pseudoobligatory, non-informative and little-informative genes are not used in genetic transformations. Effective algorithms for the synthesis optimization were realized in the intelligent recognizing system, which contains the procedures of selection of able-to-compete individuals from populations, decision making concerning the object under investigation on each able-to-compete individual chosen from populations at the second stage, and organizing of voting on a set of these individuals.