ARCHITECTONICS OF CALCULATED MATHEMATICAL MODELS UNDER UNCERTAINTY

Abstract

This article concerns the improvement of calculated mathematical models of technological, biotechnological, and economic systems. It is necessary to increase the number of considered parameters to increase the accuracy of calculating the parameters of complex systems during mathematical modeling. This leads to the need to solve nonlocal boundary value problems with non-stationary differential equations, to prove the correctness of which it is impossible to apply the traditional theory of existence and unity of solution. Note that after the architecture of boundary value problems assumes the existence of their solution, it is only necessary to prove its uniqueness. To prove the correctness of calculated mathematical models requires neither generalizing the parameters of the goal function and using approximate constraints, which, in turn, will reduce the boundary value problem to a standard form and its correctness will not be in doubt, nor propose a method to prove the correctness of boundary value certain differential equations, which will consider the specific features of the modeled processes. A separate technique must substantiate the correctness of boundary value problems depending on the type of differential equation that describes the physical and economic processes in the simulated systems. This article studied the conditions for the correctness of boundary value problems for differential equations with constant coefficients. It is proved that there is a corresponding boundary value problem for arbitrary homogeneous differential equations. It is defined the parabolic boundary value problems in terms that use constraints from above on the fundamental solution function. The conditions were obtained under which the parabolic boundary value problem exists and cannot exist, respectively. The obtained results will increase the accuracy of the main optimization task of improving the quality of simulated processes.