Instruments and Systems: Monitoring, Control, and Diagnostics Annotation << Back General principles of selecting parameters of multi-agent stochastic search algorithms for solving specific problems of discrete optimization O.V. YESIKOV, D.O. YESIKOV, N.Yu. AKINSHINA The list of problems of discrete optimization used in practice is given, for the solution of which, due to their large dimensionality, nonlinearity of objective functions and limitations, traditional methods cannot be used. The use of multi-agent algorithms for stochastic search with an island scheme for organizing computations has been experimentally tested for solving these problems. The experimentally obtained parameters of multi-agent algorithms and approximating dependences for their determination are provided, which provide the result of maximum quality for a limited time in solving various problems of different dimensions. The possibility and mathematical dependencies of the application have been experimentally confirmed and their application makes it possible by varying the values of the parameters of the algorithms, to control the accuracy and time of obtaining the solution. Keywords: discrete optimization, parameters of multi-agent algorithms, mathematical modeling. Contacts: E-mail: mcgeen@gmail.com Pp. 28-37.