Search

Functioning of the enterprise cannot be effective without investment activity. This paper examines the actual scientific problem of developing models and procedures of investment projects portfolio formation in the context of investment decision support, taking into account the impact of many external factors and changes in investment policy. The paper shows the investment portfolio formation for a specific investment policy based on different approaches. These technologies are implemented using dynamic and Boolean programming. Dynamic portfolio formation technology, which allows to plan investment resources at both tactical, and strategic level, is proposed. Projects for each time interval are selected according to the risk and return level of investment policy. Then, iterative process of portfolio formation begins. Optimal portfolio is formed on the first interval, where return serves as a criterion. All projects that were not included in this portfolio may be transferred to the next interval in the case the client agrees to get investments on the next time interval. The client can also reject investments. In this case, the projects that were not previously included in this interval may enter the portfolio instead of rejected. The iterative process ends when the selected investment policy is implemented. Based on the data of mathematical models and technologies, the decision support system in the form of program solutions, which allows to make scientifically-grounded decisions by forming the corresponding investment portfolio will be implemented. This system can be integrated into a corporate management system of the company, engaged in investment activities. [ABSTRACT FROM AUTHOR]

The problems of decision-making under uncertainty are considered in the paper. Uncertainties are an integral part of decision-making processes in almost all organizational and technical systems. It is the practice to divide decision-making problems into well-structured, semi-structured and unstructured ones. The latter are the subject of the study in this paper. The methodology of a verbal decision analysis is regarded as the basis for unstructured decision-making problems. In combination with the known scientific approach in the decision theory MAUT (Multi-Attribute Utility Theory), the basis for implementing the verbal approach to decision-making serves the construction of a utility function, being axiomatically justified and providing a decision rule for any hypothetical alternatives. For decision-making, the mixed fuzzy evaluation method, which allows in contrast to the known methods avoiding a time-consuming procedure of pair-wise comparison of alternatives, is proposed in the paper. Since the method is based on a verbal analysis, it allows "communicating" in a natural language with an intelligent decision-making support system [ABSTRACT FROM AUTHOR]

The architecture of agent-based manufacturing systems is considered, a conceptual model of a manufacturing agent (MA) is described. The required operation features of manufacturing agent systems is self-learning, self-organization and self-adapting. The purpose of multi-agent systems is achieved by the interaction of manufacturing agents. In practice, it is offered to implement a manufacturing agent as an intelligent robotic system, which can move within the manufacturing area, carry out transport, auxiliary and separate manufacturing operations. One of the peculiar features of the MA operation at the workspace is the determination of the process equipment position and indetermination of the transport system position, other agents and humans. Therefore, a peculiar feature of the manufacturing agent control system should be the adaptability of a decision-making system, which can create and, if necessary, modify plans of the MA operation in conditions of changing the workspace of a flexible integrated manufacturing. It is offered to observe the MA workspace by using advanced sensor systems. Herewith, one of the methods of setting assignments for the MA can be the setting by visual methods and thus, the implementation of the MA visual control. It should take into account the possible multi-zone nature of MA workspaces. The paper results were used in developing the intelligent robotic system of assembling [ABSTRACT FROM AUTHOR]

The paper proposes a solution to the patients' differential diagnosis problem by a combined application of the basic game theory, linear programming, and the conjugate gradient method. Based on the clinical research, a differential diagnosis of diseases such as atopic dermatitis, dermatitis herpetiformis, drug disease, psoriasis, eczema, scleroderma, lupus erythematosus, mycosis in 590 patients was carried out. Serum levels of T-lymphocytes, T-helper lymphocytes, cytotoxic T-cells; B-lymphocytes; s5 - interleukin 2; interleukin 6, endothelin-1 polypeptide; the percentage of segmented neutrophils in a leukogram; the serum IgA level; the se rum IgM level; the serum immunoglobulin G level, s12 - the serum total immunoglobulin E level were considered informative features. As a result of applying the proposed method not only the patient's diagnosis was determined, but also the doctor was given the opportunity to weigh the probability of possible medical diagnostic errors. Using the approach, proposed by the authors to differentiate diagnoses with similar clinical implications, allows improving the efficiency of medical decision-making. [ABSTRACT FROM AUTHOR]

The problem of assessing the risk, arising from implementing the enterprise procurement plan is considered in the paper. The task of forming the procurement plan is reduced to solving a non-trivial problem of rational allocation of total investment between private investment projects, which together determine the selected procurement strategy. This problem is formulated as an optimization problem of mathematical programming with nonlinear separable objective function and linear constraints. The allocation efficiency criterion is profit, expected from implementing the procurement plan. The real difficulty in solving this problem is that the parameters of the objective function are random variables. Therefore, for any allocation of investment funds, which satisfies the constraints, the allocation efficiency criterion value also becomes random. In this connection, there is a specific problem of estimating the uncertainty level of the result of funds allocation between investment projects and the risk of possible solutions. The problem is solved as follows. First, statistical estimation of probability-theoretic characteristics of the objective function parameters is performed. Furthermore, by expanding the target function in a series, its linearization is made. After that, the calculation of the mathematical expectation and variance of a random value of the criterion is conducted that allows to determine interval estimates of the expected value of the profit, expected from implementing the selected procurement plan. It is clear that the obtained plan efficiency estimates are more informative than point and can be used when selecting the rational investment plan. [ABSTRACT FROM AUTHOR]