Analysis of existing approaches to setting the intelligent management systems of transport undertakings

Abstract

The problem of designing intelligent management systems (IMS) of dynamically variable objects (DO), operating under significant a priori uncertainty is considered. The analysis of existing approaches to developing DO IMS, methods, models and algorithms of their construction based on the integration of classical methods of management theory and artificial intelligence methods was presented. As examples of DO, rolling stock (TU) of the multi-mode enterprises is examined. The range of unresolved problems is identified, the purpose and objectives for the solution are formulated. Currently, the problem of designing the automatic management systems of dynamically variable objects is characterized by the transition from the paradigm of adaptive management to the paradigm of intelligent management. This is caused by a continuous complication of management objects and conditions of their operation, the emergence of new classes of computing means (in particular, distributed computing systems), high-performance telecommunications channels, and a sharp increase in the reliability and efficiency requirements for management processes under significant a priori and a posteriori uncertainty. Accounting of these factors is possible only on the basis of transition from "hard" algorithms of parametric and structural adaptation to the anthropomorphic principle of management formation.