One of the main tasks of the decision support theory is the study of methods and tools for solving the problem of minimizing the negative consequences and risks in choosing strategic directions for the development of the studied system - the object of management. The formal algorithm of the optimization in conditions of the decision-making process stochastic uncertainty, and realization of steady states in system is investigated. The purpose of the algorithm model is to provide the predicted dynamics, compensation of structural, parametric uncertainty of the control system. The ambiguity of the choice the alternative solutions and as a consequence - the inadequacy of the mathematical model, due to the significant amount of stochastic and functional relationships, different ways of presenting input data, the impossibility formalizing the studied processes. Solutions in conditions of partial or complete uncertainty can be found by searching for elements of a set the alternatives, each of which with some probability may be the optimal solution. If statistical observations of the studied object or management process are incomplete, insufficiently formalized, or impossible at all, then the uncertainty of the decision to predict the directions of their possible development is clear. The decision-making process in conditions of uncertainty is proposed to be divided into stages: specification and formalization of the decision-making model; choice methods and algorithms for constructing alternatives taking into account the peculiarities of the chosen decision-making model. Parametric uncertainty is described as an interval estimate of possible values of the studied parameter. The interval can be strictly limited by numerical values, or with not clear limits - descriptive qualitative variables. Modeling of the control process in conditions of stochastic uncertainty is based on the definition of the object under study as a complex system. A promising area of research on this topic is a mathematical description of the value distribution function within the interval, which can be formalized on the basis of expert estimates, or as a heuristic probability distribution function of unpredictable events.
Query Minimization under Stochastic Uncertainty
