Methods for solving inverse problems of economic analysis

Abstract

Ekaterina Borisovna Gribanova - Tomsk State University of Control Systems and Radioelectronics.

E-mail: katag@yandex.ru

The inverse problem answers the question "how do I...?" and the purpose of solving such problems is the formation of optimal management decisions. The article presents methods for solving inverse problems of economic analysis. The paper describes existing methods: inverse computation, nonlinear programming, and methods developed by modification of known. In particular, it is proposed the use of regression functions and factor of influence. In the first case, the researcher generates a linear regression function of the dependence of values, which are arguments, and solves the resulting system of equations. The use of regression allows reducing the number of checks match the original data set objective and requires determination of direction of dependence: direct or inverse relationship between the arguments. As an example, consider the solution of the inverse problem for the multiplicative two-factor model. There is an example of an additive three-factor model and the problem solving, the application of the convolution procedure and of the system of equations. The method of factor effects based on the theory of factor analysis, in particular on equity total changes of result and amount of changes the magnitude of the result due to each factor. The remainder of the interaction of factors is distributed evenly between the arguments. The study used the methodology of the theory of inverse computation, the minimax method of estimation regression parameters, methods of factor analysis, optimization methods. The specialist selects the method depending on the requirements to the obtained results, availability of inputs coming from the decision makers, software. System of support of managerial decisions can use the methods developed.