Creation principals of universal modeling environment for simulation application development


            Creation principals of universal modeling environment for simulation application development


S.A. Vlasov*, V.V. Deviatkov**, T.V. Deviatkov***

*Department on nanotechnology and information technology RAS, Moscow, Russia
(Tel: +7-495-938-18-91; e-mail: savlas@yandex.ru).
**Elina-Computers Ltd., Kazan, Russia (Tel: +7-843-236-19-94; e-mail: vladimir@elina-computer.ru).
***Kazan State Technological University named after Tupolev, Kazan, Russia
(Tel: +7-843-236-19-94; e-mail: the9th@yandex.ru).


            Abstract: The structure principles of the universal modeling environment and it's means of different simulation application arrangement are described in the report. There are examples of environment usage according to separate application realization. The perspective of it's development is shown.
            Keywords: Simulation modeling, simulation research, simulation application, program modules, universal modeling environment.

1. INTRODUCTION

The usage of Simulation Modeling in practice is not just a model creation, but it is a problem definition, data collection and accumulation, experiment plan development and it's realization, detailed analysis of results and etc. (Neylor, 1975).

Thus in practical usage of Simulation Modeling it is require not only set of program means presentation for automated execution of certain periods but the presence of unified and integrated modeling environment on the base of which can be formed quickly simulation application for simulation research in a chosen subject area (Deviatkov, 2005; Vlasov, et al., 2007).

Currency of creation of such modeling environment is rising last time so as there is a great demand for the practical simulation researches. It is confirmed by authors' practice so as if several years ago the talk about simulation cause a smile but now in most cases it is point of start work which depends on only if a user has resources for this work.

Also it is necessary to point out that created and run instrumental devices of simulation modeling (common purpose languages, commercial simulator and etc.) (Vlasov and Deviatkov, 2005a, b) don't let make the simulation research process from "art and science" (Shannon, 1978) into an engineering task accessible for many system's analysts. Nowadays the simulation modeling usage is not widespread in a real economy and that is why only few specialists can use this instrument.

That is why it is necessary set flexible and unified software which help to develop fully functional and easy in usage simulation application on a tight timetable for a definite subject area.


2. CONCEPTS AND FUNCTIONAL STRUCTURE OF THE UNIVERSAL MODELING ENVIRONMENT

Principles of simulation application creation are the basis of the universal modeling environment which are described authors in (Vlasov, et al., 2007). Methodologically everything is based on some stages automatization of simulation research and then joins them into a single program system - simulation application.

In view of the fact that there are unlimited detail means in simulation modeling, capacity and power of modern computers and modern program technologies development the simulation application becomes the synthesis of classical simulation modeling stages and other information technologies. It can include also necessary for research analytical and physical models as way of simulation object presentation, data filling and analysis of studying system (Fig. 1). It was clear by intuition before but it is more obvious now in globalization and integration era of all systems and processes.

Analytical models are all possible algorithms and calculation methods, control and data, forms and presentation ways input.

Physical models are usage if it is necessary of life testers, simulators and other system prototypes and its elements in the process of simulation experiment.

The authors had the idea of simulation environment's creation many years ago (in 2003) and it is realizing step by step.

It is necessary to point out that modeling environment is only "semi-finished product" that is it is not fully ready program product for analyst’s usage but it is only set of construction elements and constructions for its creation. It is not only software but also joining methods, standard interfaces and other procedures. In order to construct from this application

the programmer must work in creative union with analyst.
Analyst should raise the question for the application in a professional way, define enumeration and database sources and develop entrance data input and results presentation forms in accordance with established terminology, traditions and document's circulation in this subject area. Also the analyst should control the adequacy of put model in the application.
Programmer should collect and customize of having means in correct way and if it is necessary improve or develop new program modules.
The universal modeling environment is creating for quick


development of simulation application by analytics and programmers. That is why it includes not only software. The universal modeling environment consists of the following parts:

- program modules, which are used as a single program complex of application;

- usage and program documentation of developer for each program modules;

- operating and architectural standards for developer with ways and rules of program models' connecting in a single integrated program system.





                Simulation Modeling





     1 - Task definition
     2  - Collecting, processing and database inputting
     3  - Modeling core
     4  - Experiments planning and realization
     5  - Results analysis, recommendations producing



■ Imitating

О Analytical

О Physical


              Fig. 1. Integration of Simulation Modeling stages and different model types in simulation application.

3. PROGRAM MODULES

Functionally all software of the unified modeling environment can be divided into a row of autonomous program modules, which have definite task. In general program module’s composition of modeling environment is closer to simulation researches process structure. But there are differences so as some program modules are universal


and can be used on different research stages or several program modules can be used at once on one stage. Also modeling environment can include analytical calculation modules, modules of connecting with physical models and etc.

In principle program module composition can be different according to application. But among plenty of pragram groups only three standard groups (subsystem) types can be

singled out, which must be included in application.

1   Program module's group of database input;

2   Program module's group of experiment organization;

3   Program module's group of results analysis.

Program modules composition in a group can be changed. For example, networking organization module is not used for local application but it must be used for distributed application.

From the point of view of ready degree of usage in simulation application creation all program modules can be divided into three parts:

1   universal program modules, which are part of simulation application without developer's changes;

2   customizable program modules, which depend on subject area and research's tasks. This dependence in a definite simulation application is expressed due to simple customization proceedings, script development and etc.;

3   developing program modules, which define exceptional peculiarities of given application.

Universal program modules. They don't require any program improvement or customization. It is only necessary to organize preliminary preparation and correct database input (hand or automated), patterns and primitives in the universal module. Also it is necessary to know output data structure, which is entrance data for other program modules. As an example of the universal modules can be 2D and 3D modules of entrance data visualization and output data animation, modeling nucleus (using common destination simulation modeling system), experiment's results and experiment's series database.

Customizable program modules. They are ready for usage but they require customization. Customization is a choice of a path to program, input of varied parameters values, library and source data files connection, definition of entrance and output data forms, choice of observation intervals and etc. As an example of program modules of this type is dynamic monitoring subsystem, which is used as a customization of data enumeration collected in the process of experiment.

Developing modules. It is more indefinite and informalized program modules type. Any subject area has nuances and peculiarities. That is why a typical example of these modules are modules, which are realizing user's interface. Taking into account that one of the most general principals of simulation application development is maximal approximation of user's interface of application cooperation of analyst in the real life and as a result there is necessity of some new program modules development.

4. OPERATION AND PROGRAM DOCUMENTATION FOR SIMULATION APPLICATION DEVELOPERS

After primary development and functionality control in a whole number of real simulation applications a program


module can be include in the universal or the customized program modules of the modeling environment. Precondition in this case is documentation of two types, which are correctly developed. The documentation for a programmer -"Programmer's Guide", "Program’s texts" and the documentation for given module usage in simulation application - "User's Guide", "Test Example".

Extremely important condition is presence of source modules, only in the case of absolute necessity a program module without source modules (without the document "Program Text") can be included in environment. Such cases are the following:

1    it is the simple in usage, widely known and tested program, which is supporting by the developer;

2    when purchase of the source program modules is very expensive or impossibly, but you have long-term contract relationship of cancelling and defect's elimination with the developer.

Examples of the first case are usage in the simulation application of such well-known programs as MS Excel and Macromedia Flash (together with the inner language ActionScript).

MS Excel was used: it has wide possibilities of table data processing; and also any user has skills of working in it; there is a necessity of quick development completion.

Macromedia Flash Program made it possible to solve questions of 2D animations construction as soon as possible. Plot's quality and simple usage of this program is well-known.

Programs, which are extremely complex and they require more financial and time resources are reffered to the second case. Besides, these programs are "de facto" standart in this knowledge area. As an example of this program is our usage of general designated modeling system GPSS World (Vlasov, et al., 2008). We used this system in all developed simulation applications, for example, (Kovalev, et al., 2007; Kozlov, et al., 2007; Samoilov, et al., 2007) as "the modeling nucleus".

5.     METHODOLOGICAL AND ARCHITECTURAL STANDARDS OF PROGRAM MODULES CONNECTION IN THE MODELING ENVIRONMENT

Presence of worked and well documented programs does not let effectively organize the process of application's simulation development. It is necessary to define the common principals of application architecture and program connection standards of all modules.

Conceptually the following basic principals of simulation application architecture can be formulated:

1    The single managering module must be in every simulation application. That is program module, which is used as a start and finish of application work and it is an interconnection of all modules, which are used in the application;

The application must create database of models and results, which has the common composition structure;

3    The application’s dialogue language of data input and result’s analysis must be as much as possible corresponds to subject area's language of research and terminology, which are established in this modeling environment (the document "Terms and definitions of Modeling Environment");

4    A user must be free from model's development on simulation modeling language. Simulation application must provide automatic generation and model's collection according to input data of user.

5    Informational connection of used modules must be maximal unified. It can be achieved by using: well-known


languages of data structuring (for example, XML); corporative and international agreements and standards of data exchange.

According to the first principal and program modules composition the simulation application architecture can be presented in the following way (Fig. 2).

In spite of that fact that there are essential distinctions in composition data representation forms in the modules "Source Data" and "Modeling Results" in every simulation application it is necessary to unify the structures of these data on the whole for all applications in accordance with the 2nd principle. The general architecture of the informational application is showed on the Fig. 3.


Fig. 2. General architecture of simulation application composition.

Fig. 3. General architecture of the application's informational model.

It is necessary to point out that this conditional structure shows the unification possibilities of application's informational model data composition. All is based on such terms as model, model's parameters, experiment with a model, and experiments series with a model. Hierarchy of these objects in the process of data model construction is fixed. There can be loose interpretation of other data structure's elements but little by little the whole hierarchy of data model will be supplemented with new details.

One of the least formalized principles of simulation application development is the 3rd principle - the user's language of data input and result’s analysis. The reason of it is striking differences of subject areas' language of every language in the application. The terminology of railway transport and its information representation is one and it is different for banks indeed. Nevertheless it is planned to create programs of computer-aided design of window graphics dialogues in accordance with the user's scripts.

The 4th principle in the universal modeling environment is realizing in every simulation application due to codes generator program, which is unique for each application. Now the subsystem of system's formal graphics description and codes generation is in the process of developing. There is a special mathematical apparatus of abstract description of complex discrete systems in this subsystem so the automatic code generation of model in the GPSS World language is carrying out in accordance with the user's description.

The 5th principle is realized partly in this universal modeling environment version. In number cases the data structuring language - XML language is used, in some cases the simple TXT files can be used. Now there are not the common rules and standards. But there is necessity to realize this principle in the near future.

6. THE MAIN RESULTS AND DEVELOPMENT PROSPECTIVE

The modeling environment work was organized in the following way, at first (2003-2007) the program modules (dynamic monitoring, experiment planning, server of distributed modeling and etc.) were realized or basic elements (modeling nucleus) were chosen from the finished programs, on the based of which can be created the simulation application. Now the real simulation application on the existing basic is developing, at the same time there is development of systematic addition of the environment's universal program means.

At present the modeling environment helped to develop a whole number of rows of simulation application for example, (Kovalev, et al., 2007; Kozlov, et al., 2007; Samoilov, et al., 2007). All of them are accepted in exploitation by customers.

Now any programmer can develop a simulation application for several months if he has a definite research task and he uses modeling environment's tools and all descriptions. It took more than 6 months of programmer’s work with

advanced qualification several years ago.
At this moment it is the main result of the usage of the universal modeling environment.

7. CONCLUSIONS
In conclusion we define the tasks, which are necessary to solve in the near future to improve the previous developments from the point of view universality.
1.  Complete the development of customizable subsystem "Formal Graphics Description of System and Codes Generation"
2.  Create a single unified model of simulation application's data
3.  Develop the universal subsystem of two- and threedimensional animation modeling results.
4.  Document and begin to use in practice a number of corporative standards of architecture, functions, program modules connection and other questions of simulation application creation.

REFERENCES
Deviatkov, V.V. (2005). Application Development in GPSS World environment. In: Proceedings of the 2ⁿd Russian theoretical and practical conference on Simulation Modeling and it's usage in science and industry "Simulation Modeling, theory and practice”. Snt.-Petersburg, Russia, vol. 1, pp. 186-190.
Kovalev, V.S., Usanov, D.I., Tsutskov, D.V., and Kalinin, S.V. (2007). Enlarged Model of Railway Direction. In: Proceedings of the 3rd Russian theoretical and practical conference on Simulation Modeling and it's usage in science and industry "Simulation Modeling, theory and practice". Snt.-Petersburg, Russia, vol. 2, pp. 91-95.
Kozlov, A.N., Deviatkov, T.V., and Kejer, P.A. (2007). Research of Functioning of Collective Information's Processing Centre by Simulation Modeling Method. In: Proceedings of the 3rd Russian theoretical and practical conference on Simulation Modeling and it's usage in science and industry "Simulation Modeling, theory and practice". Snt.-Petersburg, Russia, vol. 2, pp. 96-100.
Samoilov, V.V., Vlasov, S.A., and Deviatkov, V.V. (2007). Simulation Research of Maintenance Services of Program and Technical Means of "Tatneft" Company. Automation in industry, No 4, pp. 11-14.
Neylor, T. (1975). Machine simulation experiments with economic models systems. Moscow, 500 pp.
Shannon, R. (1978). Simulation Modeling of Systems - Art and Science. Moscow, 418 pp.
Vlasov, S.A., and Deviatkov, V.V. (2005a). Simulation Modeling in Russia: the past, the present, the future. Automation in industry, No 5, pp. 63-65.
Vlasov, S.A., and Deviatkov, V.V. (2005b). Experience in the development and application of simulation in Russia: review, analysis of prospect. In: Proceedings 19th

    European Conference on Modeling and Simulation. ECMS 2005, June 1-4, pp. 23-27.
Vlasov, S.A., Deviatkov, V.V., and Kobelev, N.B. (2007). Methodology, technology and program realization principals of simulation application. In: Proceedings of the 3rd Russian theoretical and practical conference on Simulation Modeling and it's usage in science and

    industry "Simulation Modeling, theory and practice". Snt.-Petersburg, Russia, vol. 1, pp. 17-26.
Vlasov, S.A., Deviatkov, V.V., and Deviatkov, T.V. (2008). Development's Conception and Methods of Distributed Simulation Applications with using of GPSS World Environment. In: Theses of International Scientific and Technical Conference "Information Technologies and Mathematical Simulation of System". Majorca, Spain.