V. A. Zelentsov, E. G. Tsivirko

Maintenance control of complex spatially distributed systems due to the need to analyze large amounts of information about the technical parameters of elements of these systems, also about functional interrelation of elements, about conditions of their application both available material and manpower resources. Fully centralized collection and processing of this information is often either technically infeasible or lead to decisions on inaccurate or outdated information, ie in conditions of considerable uncertainty. The output is a partial decentralization of decision-making, thus the general problem decomposed on a number of the subproblems solved by the upper level control center and the lower level control elements (CE) which are approached to sources of the necessary information. Decentralization leads to a reduction of uncertainty about the technical parameters and operating conditions, but it leads to another kind of uncertainty associated with the independent actions of CE. CE interests at decision-making may not coincide completely with the interests of the control center and whole system. The permission of arising contradictions is reached by attraction of methods of the games theory. Use of these methods and their development with reference to features of maintenance control problems has allowed: To formalize the problems of the control centre and CE; To establish that coordination of the decisions accepted by elements of hierarchical system, is reached by developed procedure in which for each element the local multicriteria optimizations problem is designed, and the optimum decision for all system is defined within of Pareto - optimum decisions of local problems. To define conditions agreed upon interests of the elements of considered hierarchical system and the requirement to the necessary information exchange between CE and the control centre.