The National Science Fair

Multiple Document Interface

Once this plan is found, viewing its components. If there is no component of fractional numbers, the plan is found optimal plan solution of integer programming. Otherwise, the system of equations is added to the transformed variables, inequality, taken from the last simplex table. These actions repeated until until it finds the optimal plan for the task, or set its insolubility. The solution of mixed integer linear programming problems is sequential decision tasks, each of which is obtained from the preceding by the introduction of additional restrictions. Additional restrictions are called proper pruning, which can be interpreted as a hyperplane, which is just like "Cut off" from the region of feasible solutions noninteger optimal solution.

In the design used the principle of modular programming, which simplifies the debugging of a program and allows you to extend its functionality opportunities. The algorithmic part of the program has a modular-hierarchical structure in which each module is a separate part of the program and interacts with other modules in the manner prescribed developers. Methods for solving linear optimization, implemented in the software and algorithmic complex, based on the construction of the simplex teblits, so the structure of the program, all algorithmic modules related to module that organizes the solution of the linear programming simplex method. The inputs to this module is the target function, indicating the type of extremum (maximum or minimum) and the restrictions imposed on controlled variables. Limits are specified in the form of equations or inequalities. From there, the second module, which formed the initial feasible basic solution.

Second, third and fourth units at each iteration method is implemented by causing the module construction of the simplex table, which they transmit the current result. Communication between modules is organized via the external data structure. For example, to specify a linear optimality criterion, the vector of controlled variables, the constraint vector and the constraint matrix used by one-and two-dimensional static arrays, and the simplex table in computer memory is represented as a two-dimensional a dynamic array that can change its dimensions, removing or adding rows and columns of the simplex table. Consider a feature of the functioning software system. For dialogue with the user used standard graphical user interface Windows, built on the library of visual components VCL (Visual Component Library), ships with Delphi. In developing the program used by MDI-technology (Multiple Document Interface – Multiple Document User Interface) that allows the user to work with several problems of linear programming. The program features an active form of dialogue that allows to choose modes: calculation, viewing and editing information, getting help, etc. The main menu contains the following items: File, Edit, View, calculating, a window reference. All main menu items cause the submenu. At the beginning of the program Some items are prohibited and are permitted only when selecting other menu items (for example, "Edit" menu, "Calculations," etc.). The program provides the ability to configure parameters of the problem: maximization or minimization, the choice of options enables you to view intermediate results of iteration, limiting the number of iterations, setting the dimension of the problem, etc.