## A.A. Tselykh ANALYSIS OF INFORMATION STREAMS REPRESENTED WITH FUZZY HYPERGAPHS
Problem of optimal choice of information services providers in Internet is connected with time and financial limitations. Proposed method of the development of decision-making support system gives an opportunity to formalize, classify and research information streams according to the demands of users, reflecting their subjective conception of “ideal” information provider. Optimal strategies of behavior are worked out based on the analysis of user demands and independent expert estimation of different qualities of information providers. Fuzzy sets, graph and hypergraph theories are applied.
Decision-making procedure comprises two basic stages. First, building the structure of researched processes to describe them with fuzzy hypergraphs. Then, analysis of various hypergraph properties to use in the applied tasks solution.
On the first stage we represent user evaluation of necessary quality attributes of information with the fuzzy adequacy , where is the set of users of information, is the set of attributes, is the adequacy graph. The results of expert estimation of the properties fulfillment by information providers is represented with the fuzzy adequacy , where is the set of information providers, and is the adequacy graph. Composition of fuzzy adequacies is built to get the grades of provider *z*_{k} preference by user *x*_{i}. As a result, final matrix *Q*_{Ф }represents the incidental matrix of fuzzy hypergraph .
Classification of users according to the preference of information providers is done through the matrix of crossing of classes of preference, where classes themselves are described with the following level set.
for every .
This method was used by author for the analysis of the situation on the market of information services, provided by Russian electronic periodicals.
Another acute problem is the task of optimization of information provider actions through the determination of users belonging to classes of several providers at the same time. It is considered expedient to unify efforts of providers by minimization of the number of cognate users and maximization of their general number. The decision adds up to the determination of minimal coverings in the hypergraph. Author continues to explore defined science field to propose an effective algorithm of minimal coverings determination.
## A.P.Suchkova DECISION SUPPORT SYSTEM FOR PUBLIC PROCUREMENTS AND TENDERS
A new Decision Support System (DSS) for public procurements and tenders is proposed. The goal is to improve the quality and the validity of decision making in non-profit and public organisations. A fundamental conflict lies at the root of any decision-making: multiple human needs versus limited available resources.
The DSS is build on the theory of analytic hierarchy process (AHP), the fuzzy sets theory, the measurement theory and improved by a new scientifically valid method for control through AHP proposed in this papers. The method belongs to the area of multicriteria decision making.
The general concept of the proposed method is to construct more than one AHP with the same set of alternatives, using different ways: strategical, tactical and operational. This approach extends decision-maker’s capabilities and allows to investigate the set of alternatives more comprehensive. The other idea is the ultimate using all kind the information: qualitative and numerical, expert judgements, as the data transformation is accepted by developed mathematical method. The intangible factors measurement is available. Unique data transforming brings a new viewpoint and arranges decision maker’s notions. Group decision is supported without using weighted average.
The latest version of DSS-software is implemented like a database with build-in DSS-tool. As data keeping before and after transforming is guaranteed the wide range of the traditional data processing methods are applicable.
## S.I. Yusifov FUZZY OPTIMIZATION OF WORKING AGENT DISTRIBUTION ON GAZLIFT OIL EXTRACTION
The distribution of the limited resource - consumption of the working agent between gaslift wells is carried out, as usual, in conditions of inaccuracy of a initial data. These data are obtained from of the exploration investigations. As a result of these investigations the parameters of characteristic models of wells are determined insufficiently precisely. It deviates uncertainty of conditions of optimization of modes. This cannot be considered as stochastic, in view of absence of the appropriate parameters. Therefore, they can be characterized by fuzzy categories. These categories depend on the given expenditure of the working agent and parameters of characteristic modes of operations of gaslift wells. The fuzzy parameters of characteristic modes of operations of wells are obtained on the basis of the solution of a task of fuzzy identification. As result of this identification fuzzy parametersare determined. Identity functions for these parameters are determined as follows:
For a solution of a task of fuzzy optimization of distribution of the working agent, the method of consecutive approximation in space of strategy in fuzzy conditions, with the account of the gradient method of movement toward optimum is used.
Therefore, in the beginning, on each well the values of derivatives are determined. Whereas both _{1} and _{2} are fuzzy coefficients characterizing the basic characteristic of a well, P_{i}, also will be fuzzy numbers. Thus, each mode will be characterized by three fuzzy numbers V_{k}, Q_{k} and .
The deduction of the fuzzy solution begins with the descriptions of the current condition of the object. For this purpose we shall consider the system made up of n gaslift wells. The current condition of each well, as was mentioned, is described by three parameters: V^{T}_{K}, Q^{T}_{K} and P^{T}_{K} where . In the general description of all system in addition enter the total consumption and total fuzzy debit .
Assume that due to the change V_{0} - of the given limit of the consumption of the working agent there arose a necessity to a task of distribution of the working agent in order to maximize the total debit.
**Share with your friends:** |