Molodih Aleksandr Gennadievich

Summary on the final work: Research and development of the topological analysis of technological schemes as for modeling

Goals and objectives that must be determined:

The objective of this work is research, development and implementation of subsystems of topological analysis of technological schemes, which should operate successfully in a distributed parallel environment simulator (DPES).

Relevance and motivation of the topic:

Simulation modeling of technological schemes using parallel modeling environments is an actual line of development, but the tools to implement these tasks, at the moment, have significant drawbacks. This work involves building a better instrument for solve current simulation problem.

Expected Scientific novelty:

Scientific novelty is in the realization and implementation of the module DPES, and in development of efficient algorithms which enables automatic parallelization of simulation models.

Expected practical results:

Software implemented subsystem of a topological analysis like a segnent of the DPES.

A review of research and development on the subject in the world:

MatLab [6] - high performance environment of mathematical modeling, which includes calculations, visualization and programming in a comfortable environment with a great set of tools. Typical tasks that use MatLab, are: math, algorithm development, simulation, data analysis, scientific and engineering graphics. MatLab works with MIMD systems at various levels, and is able to use technology CUDA.

Simulink - MatLab accompanying interactive program, which takes possibility to block modeling nonlinear dynamic systems.

GNU Octave [7] - high-level mathematical modeling environment. The main purpose is in using numerical methods for solving linear and nonlinear equations and other mathematical experiments, and their subsequent output of numerical and graphical format. GNU Octave is open source program, and can be ported to some of the hardware.

Scilab [8] - a package of applied math programs, which contains a powerful open environment for engineerc, scientific and technical calculations. This system contains tools for working with 2D and 3D graphics, animation, linear algebra, sparse matrices, polynomial and rational functions, differential equations, signals, and many other professional tasks. Scilab is open source program and can run at different operating systems and hardware.

Scicos - a tool of Scilab for editing block diagrams.

Summary of individual results which availables at completion of the abstracts.

Simulation modeling of technological schemes which using parallel modeling environments is an actual line of development, but the tools for carrying out this tasks have significant drawbacks at the moment. Subsystem TA TS DPES is a step to more perfect tool of simulation. The purpose of this report is in detailed consideration of development the subsystem TA TS DPES, which contains in analysis, solutions and technologies that should be applied to developing subsystem.

The main stages of subsystem TA TS work

The scheme of subsystem TA TS work is shown in Figure 1.

An input data for the subsystem of topological analysis technological schemes is in the description topology file. When dialogue subsystem is receiving input file, the subsystem of the topological analysis begins processing the file for errorsafter, this stage is in check for the basic matrix connections at each level of the hierarchy, check for TS element information at the common matrix (graph), chech for the bonds ang other things.

These stages of the TA work are associated with periodic calls to library elements topologies, is done to confirm availability in the database components which are used, and obtaining the connection matrices which associated with the element used in cases where the item is not a representative of lower-level component hierarchy.

When common bond graph is formed, subsistem must decomposite it into groups of elements. Each group represents the amount of computation falls on one process - the granules. Grouping is carried out by the following criteria:

- must aspire to minimize the external bonds between groups;

- the groups count should be depend of available resources;

- computational intensity in each group should be strive for uniformity.

The last condition is difficult to achieve, because different base elements have different computational intensity. As a solution of this problem is in attaching to every basic elements some abstract values of computational intensity specific. This value can be entered into the database manually during the creation of elements - the static evaluation, or calculated by the software - a dynamic evaluation.

The final stage of subsystem topological analysis work is in transfer the results of the TA in the dialogue subsystem. The result is a common bond graph and parallel model. Parallel model is a list that contains elements of the total bonds graph.

Resolution: Vector Graphics

File Size: 64.4 KB

Duration: 13 sec

Palette: 24 bit

Fig. 1. The scheme of the subsystem of the topological analysis of TS

In [1] to work with a common bond graph proposed to use a matrix. But whan simulation of real circuits is being, it represents by thousands of elements, that may be some redundancy in the number of iterations at algorithms processing and memory use. To solve this problem is suggested to use other ways to represent graphs, such as a list of edges. Final approval of the storage method for graph is need to analysis performance of algorithms inherent in each method of storage.

Conclusion

In this topic is analised and elucidated main principes of subsistem TA TS development which were in other topics.

Provided new principes and solutions which takes grow up subsistem performance.

Provided solutions which permits reducing of erors that are have place at development and implementation subsistem.

Literature

1. Бондарева К.С. Топологічний аналізатор технологічно орієнтованого паралельного моделюючого середовища / В.А.Святный, К.С. Бондарева // Наукові праці ДонНТУ, серія МАП. – Донецьк, 2005. – №78. – С.5-12.

2. Святний В.А. Паралельне моделювання складних динамічних систем // Моделирование – 2006: Международная конференция. Киев, 2006 г. – Киев, 2006. – С. 83–90.

3. Бондарева Е.С. Объектно-ориентированный подход к параллельному моде-лированию химико-технологических процессов / Бондарева Е.С. // Наукові праці ДонНТУ, серія «Інформатика, кібернетика та обчислювальна техніка»–2010. – Вип.9 – С.100-105

4. Modelica [Electronic source] / Андрей Зубинский - Access mode : http:// ko.com.ua/modelica_34622, free. – Title from screen. – lang. rus.

5. Wikipedia [Electronic source]. - Access mode : www/ URL: http://ru.wikipedia.org/wiki/Обработка_исключений. – Title from screen. – lang. rus.

6. Реферат Математика Matlab [Electronic source]. - Access mode : http://works.tarefer.ru/50/100144/index.html. – Title from screen. – lang. rus.

7. Octave [Electronic source]. - Access mode : www/ URL: http://www.gnu.org/software/octave/. – Title from screen. – lang. eng.

8. Home - Scilab WebSite [Electronic source]. - Access mode : www/ URL: http://www.scilab.org/. – Title from screen. – lang. eng.

9. Буч Г. Объектно-ориентированный анализ и проектирование с примерами на C++ / Буч Г. – [3-е изд. / пер. с англ.] – М.: «Издательство Бином», СПб.: «Невский диалект», 2001. – 560 с.

10. Святний В.А. Паралельне моделювання складних динамічних систем / Святний В.А. // Наукові праці ДонНТУ, серія "Проблеми моделювання та автоматизації проектування динамічних систем" – 2006. – вип.5 – С. 5-20.

Notice.

During writing this abstract the master work was not finished yet. Final completion: December 2011. Full text can be obtained from the author or his teacher аfter that date.