Про ідентифікацію параметрів у задачах оптимального інвестування

Abstract

Розглядається задача про побудову оптимальних значень параметрів математичних моделей, що застосовуються при інвестуванні у цінні папери.

This study aims to address the problem of identification of parameters of mathematical models of dynamic processes that can be described by ordinary differential equations and systems of equations. In this paper we outline three different approaches to build optimal parameters of parametric models of dynamic systems. On the example of mathematical models of dynamic formation of the market value per share and stock portfolio developed algorithms for constructing optimal values of the parameters of such models. The algorithm can determine the multiple parameters estimation of mathematical models of dynamic processes in the class of ellipsoidal sets. When solving applied problems of financial analysis it allows researchers to obtain guaranteed financial performance of investment operations. The algorithm is based on the use of iterative procedure at each stage which formed ellipsoids containing the model parameters obtained through solving the additional problems of parametric identification. Each point in the space of such a set of model parameters is internal, or belongs to its boundary.