A fundamentally new approach to the implementation of information collection and analysis processes in the systems of operational planning and functional reliability management, based on the creation and application of analytical information systems (AIS), is proposed. The main purpose of such systems is the creation of information platforms in the interests of decision support systems (DSS - decision support system). Within the limits of AIS of the functional reliability management the creation of the unique mechanism which is development of integration OLAP- and ROLAP-technologies in the following directions is supposed: an operative change of structure (list) of the stored data and their interrelations; multistage access to the data and formation of flexible not planned queries for a sample of the required data in terms of the subject area; support of unique protocols of updating of a data storage. Due to the distributed nature of data storage and processing in AIS and the specifics of organizing large-scale data storages, the computing environment and telecommunications platform of AIS should be implemented as a local area network (LAN), equipped with telecommunications facilities for communication with external sources and consumers of information. It is advisable to form AIS architecture according to the principle of multilayer organization, distinguishing four main architectural layers: application layer; middleware layer; data layer; AIS platform.
To ensure the operation of automated control systems, a new methodology of precedent analysis is supposed to be used. The quantitative search for the situation-analogues is carried out by methods of automatic sequentially sliding review of retrospective data, presented in the form of multidimensional time series in specialized databases. Using a statistical measure of proximity between the current situation and the viewed situation, a set of quantitative precedents is selected, each of which corresponds to qualitative descriptions of the corresponding time period. The specialist obtains qualitative precedent data and selects the most plausible precedents. The situations following the precedents on the time interval form a set of precedents of the predicted situation, allowing to form a qualitatively new mathematical forecast, knowingly considering all historical experience, stored in the memory of the computer, and corrected with the knowledge and intuition of a specialist. Thus, the ideology of precedent analysis provides an opportunity to harmoniously integrate the efforts of a specialist in the field of operation and machine and to come close to the formation of hybrid intelligence technologies. The proposed concept of precedent analysis does not require additional mathematical training from an operational specialist, nor does it impose exceptionally high requirements to his personal experience. The specialist can use traditional qualitative methods of situation analysis, machine search methods already selected for him. This allows to build a new type of interface, “Friendly” to the subject specialist and satisfying the above requirements.
