MODELING FOR ADOPTION OF DECISIONS AT CONTROL
Machine modeling is an effective tool for studying the characteristics of the process of functioning of complex systems at the stage of their design. But this is not limited to the possibilities of this method: in modern control systems, machine simulation is used directly in the control loop, on its basis problems of forecasting are solved for making decisions on the management of the object, i.e., adaptive control systems are implemented. The construction of such adaptive systems made it possible, on the one hand, to resolve a number of issues of the information approach to the management problem, and on the other hand, the post-processing of modeling problems in real time on modern computers, taking into account the limited resources in the object management system.
GNOSEOLOGICAL AND INFORMATION MODELS AT CONTROL
Creating a control system (CS) by distinguishing objects requires a large amount of information about the object itself, and about its input and output variables. This information is necessary for the construction of an adequate SS model, on the basis of which the management process can be effectively implemented. In this case, it is necessary to distinguish two types of information necessary for the construction and improvement of the model and SS: a priori and current. A priori information about the control object (OA), its input and output variables, internal states is necessary for constructing a model, by which the SU will be created by this object: the structure, algorithms and parameters of the CS, the criterion of functioning are selected. Usually for complex newly designed Shelters there is no model necessary for the creation of the SU, and the management task should be solved in conditions of insufficient or completely absent a priori information about the object. This is the absence of an information ("management") OU model that establishes the relationship between output and input variables [41, 43, 54].
Features of the control system.
The problem of creating an SS inevitably arises in the development of the OS and in their modernization. At first glance, it may seem that in those cases when a new SS is being developed for a system that has been operating for a long time, 5 long
the actual time in service, the position with the a priori information is better and the construction of the model is simpler. Experience shows that this is not so, and obtaining an information model is very laborious in this case. Thus, both for the case of the newly designed system 5 and for the already functioning system, the problem arises of obtaining additional information for the creation of the SS. The only effective way to obtain such information at the present time is machine modeling.
In the event that the SS is created and operates together with the system 5, managing it, there is a need to obtain current information, caused mainly by two reasons. First, it is the need to improve the SS, and second, the need to clarify the behavior of the system and the situations that arise in it to compensate for changes in the characteristics of the system 5 as an op amp. The processes with which the current information of the first type is related are slow enough and they require an evolution control subsystem to control them, and processes of the second type are faster and require a real-time operational control subsystem (RMS) to manage them.
It should be emphasized that in terms of the pace of decision-making and the place of solving problems, the subsystems of evolutionary and operational control differ significantly from each other. For example, operational control processes can be several orders of magnitude faster than evolution control processes.
The most important task of modern management theory and practice is the construction of the OU model, that is, the formalization of the laws governing the functioning of the object. On the basis of this model, the structure, algorithms and parameters of the SS are determined, hardware-software means for implementing the system are selected. One of the effective methods for constructing a model of a complex object is identification.
The broad development of work on the formalization of processes and the construction of their models in many areas of research (engineering, economics, sociology, etc.) has two main objectives. The first of them is associated with a significant increase in the ability to study on the computer base complex functioning processes of various objects using the modeling method, which requires a mathematical description of the process being studied. No less important in technical systems are the models used to achieve the second goal, i.e., applied directly in the object control loop.