When solving a wide class of practical problems, algorithms are often used that make mistakes in calculating elementary properties or refusal to solve problems. In such cases, to solve the same problem, several incorrect algorithms are usually used, and then a correcting function is constructed [1, 2]. Since the result of calculating an elementary property can be either 0 - refusal to calculate, or 1 - the property is satisfied, or 2 - the property is not satisfied, then the correcting function is a function of three-valued logic. For meaningful reasons, not many corrective functions should be constrained