“LINPARSYS” APPLICATION SUITE FOR THE A(t)mxn·X(t)nx1=a(t)mx1 TYPE PROBLEM SOLUTION

Abstract

An application suite (AS) is presented using the developed analytical and numerical-analytical decompositional methods with indirect and direct approaches for the solution of correct and incorrect problems of linear one-parametric systems of finite equations. The suite has been implemented by JAVA programming language. The MATLAB software has also been utilized through which a number of operations, functions, concepts (such as elementary operations with matrices, calculation of inverse and generalized inverse matrices, etc.) are calculated. The suite consists of 5 program-methods. For each method a separate matlab-file has been created which represents a sequence of particular orders.

The components of matrices and other required data are obtained with the help of SWING library. The user chooses a method with the help of windows, enters the matrix-related data, such as matrix size, elements of matrices and later also enters some more data related to differential transformations (for example the center of approximation, the recovering function (Taylor, Maclaurin ratios, etc.), then clicks the “Start” button which enables the call of the corresponding method with required parameters. Depending on the chosen method and data, the user receives the answer some time later, which, in its turn, is being received through the recovering function chosen by the user himself. After the data entry from JAVA to MATLAB environment the data processing takes place after which the method itself is launched. For each method a checkup operation is carried out during which the solution of the problem is compared with solutions obtained through other methods.

As allows to widely use the performance capabilities of modern information technologies and obtain effective computational procedures at solving linear one-parametric systems of finite equations.