A method for training data transmission routes in wireless self-organized networks is proposed. Features of construction of networks of this class are described. The main tasks of functioning of the control system of wireless self – organized networks are shown. The main teaching methods used to predict changes in data transmission routes are analyzed. The efficiency of application in certain fields and non-compliance with the requirements for the method being developed are explained. The essence of forecasting and direct connection with the process of learning data transmission routes is described. The routing system is shown as a necessary component for uninterrupted operation of wireless self-organized networks. The essence and requirements for the teaching method are shown. The learning unit of the forecasting subsystem is considered. Options for increasing the efficiency of scalar optimization are shown. The essence of the method is to learn the parameters (total latency; network routes; minimum bandwidth; reliability; load; load) of data transmission routes using the scalar optimization method, designed to dynamically select the most efficient adaptability function used in each new generation of evolutionary algorithms. Optimization problems with auxiliary criteria and reinforcement training are analyzed. The “Evolutionary Algorithm and Reinforced Learning” algorithm allows you to control the execution process of the evolutionary algorithm. The Hierarchical-if-and-only-if function problem is described and its efficiency when working with different algorithms is shown. The parameters used in the work correspond to the parameters of the research, which makes it possible to compare the results with previous results. In the course of work on the method its efficiency is reflected and the comparative analysis with similar methods of multicriteria optimization is carried out.