Methods for solving mathematical programming problems and statistical decision-making problems (pattern recognition problems) are presented. Algorithms are considered that make it possible to take into account the influence of errors of all random variables appearing in the problem (confluent analysis). Theoretical material is available to persons proficient in mathematics within the scope of the technical university program; real examples are considered, for example, the identification of earthquakes and weak explosions based on the results of seismic observations, the identification of aircraft, the assignment problem, the maximization of production output, etc. The textbook was created on the basis of lectures and practical classes for students of MSTU. N.E. Bauman. For students of technical universities, specialists dealing with decision-making problems, as well as students of courses in the system of additional professional education studying such problems.