The management technology related to model-based manufacturing system engineering and processing of random data is considered. This technology is based on dichotomy models and their applications. If the parameters of the random process are known, the primary dichotomy model is used to process the data. If these parameters are unknown, the self-tuning procedure is used. In secondary dichotomy, two such procedures are combined into a complex self-tuning procedure using system engineering. Such complex procedure is used to manage an agent in an environment of incomplete awareness. However, the vi-sionary agent can manipulate his indicators in such a way as to maximize own target function, which depends on the results of the secondary dichotomy. This may result in poor management. A formal method for solving this problem is proposed, based on an ontology-based management model that includes both self-tuning of secondary dichotomy and corresponding stimulation. The condi-tions are determined under which the ontology-based model increases the qual-ity of management due to the greater interest of the agent. Based on this model, data processing and analysis of the performance of renovation of railway loco-motives are carried out.