Automation of Simulation Steps using Ontological Approach

Elena Zamyatina

, Alexander Mikov

and Viacheslav Lanin

National Research University Higher School of Economics, Perm, Russian Federation

Kuban State University, Krasnodar, Russian Federation

Perm State University, Perm, Russian Federation

Keywords: Simulation, Ontology, Automation of Simulation Steps, Ontological Approach, DSL.

Abstract: There are different Modelling and Simulation (M&S) life cycle’s steps described in the literature. One way

or another some of these steps are similar: creation of conceptual model, verification and validation of

simulation model, statistical data collection and processing. Authors of represented paper suggest to automate

some steps and propose to use ontological approach. The automatization of steps allows to optimize the overall

time of simulation experiment, to increase a reliability of simulation model and to receive more adequate

results.

1 INTRODUCTION

Nowadays simulation became one of most useful and,

maybe, single method for investigation of complex

systems in various areas of business, economics,

healthcare, etc.

The design of simulation models, simulation,

analyses of the results of simulation experiment may

be fulfilled using the special software tools –

simulation systems.

It is well known that simulation process has some

steps: a development of conceptual model,

verification and validation of model, simulation run,

analyses of the results of simulation run and so on

(Balci, 1998; Balci et al., 2002; Law and McComas,

2001; Sargent, 2005; Sargent, 2007, Salmon and

Aarag, 2011).

Because researcher wishes to receive reliable and

truthful recommendation in order to make a decision

and because these recommendations have to be

received rather quickly it is advisable to optimize

steps mentioned above.

Further, we will consider the efforts of the

developers of simulation system TriadNS to optimize

and automate the steps of the simulation process

using the methods of artificial intelligence. One of

these methods is ontological approach.

2 MOTIVATION

Simulation usually deals with complex system, thus it

is necessary to have sufficient computational

resources to shorten the execution time of the

simulation experiment, since it is very important that

the simulation experiment be completed at a

reasonable time (Salmon and Aarag, 2011).

The development of simulation models requires

significant efforts from users. Most users are

specialists in a specific subject area and do not have

the art of programming. For this reason, simulation

system is required to have a convenient and user-

friendly interface. Moreover it is advisable to have

visual programming language describing simulation

model.

A useful property of simulation system is the

ability to customize the interface to a specific subject

area. Users should be able to work in the software

environment using familiar terms, operate the

construction of the modeling language (including the

visual one) (Cetinkaya and Verbraeck, 2011).

The task of the designer is to build the most

"adequate" model. “Adequate” model have to

describe an investigated object or situation in detail

and rather precisely. An “adequate” model can be

obtained using a multi-model approach (Sokolov and

Uysupov, 2005), transforming one model into another

in the course of research.

As a result of the simulation experiment, the user

often receives a large amount of unstructured data.

Zamyatina, E., Mikov, A. and Lanin, V.

Automation of Simulation Steps using Ontological Approach.

DOI: 10.5220/0006933102230230

In Proceedings of the 10th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management (IC3K 2018) - Volume 2: KEOD, pages 223-230

ISBN: 978-989-758-330-8

223

Therefore, it is advisable to provide the simulation

system with software tools for additional data

processing so that additional conclusions can be

drawn about the results of the simulation. It is

necessary to make effective recommendations and to

receive the most appropriate decision.

So simulation system must have:

1. Flexible software and language tools for

simulation model development.

2. Software tools and languages (may be visual)

for optimizing the simulation experiment in

time.

3. Software tools for verification and validation

of simulation model.

4. Software tools and languages for data

collection, processing and additional tools for

analyses (may be with the help of special

methods - data mining for example).

Special mechanisms and special data structures

are needed to develop flexible software, the purpose

of which involves automating and optimizing the

steps of a simulation experiment. The authors of this

paper attempted to use ontologies.

The paper will then be structured as follows:

simulation model in the TRIADNS modeling system,

basic ontology, ontologies automating the creation of

a simulation model, ontologies for customization on

a specific subject area, simulation model

transformation.

Let us consider several papers discussing how

ontologies may be used in simulation systems and

how ontological approach may be applied for

simulation steps automation and optimization.

3 RELATED WORKS

It is well-known that ontology is defined as a method

of representing items of knowledge (ideas, things,

facts, etc.) in such a way that determines the

relationships and classifications of the concepts

within a specified domain of knowledge (Zaletelja,

2018).

Ontologies allow researchers, domain experts,

and software agents to share a common understanding

of the concepts and relationships of a domain. So

ontologies are used in several simulation systems for

a large number of domains (a lot of publications about

application ontologies in different domains

appeared).

Thus, a foundational ontology for manufacturing

– system modelling is proposed in (Zaletelja, 2018)

Quoting the authors, we can say that “by the formal

definitions of the modelling environment itself enable

the definition of the manufacturing system’s

elements”. Ontological-based approach “ensures the

consistency of ever-changing models”.

Modeling framework based on an ontology

network is described in (Sarli, 2005). This framework

is created to conceptualize supply chain (SC) and

simulation domains, besides through the execution of

derived axioms, integrity axioms and rules in the

ontology network the composition of the SC model is

validated.

A methodology and applications of ontology-

based simulation are presented in (Beck et al., 2010).

An environment for building simulations based on the

Lyra ontology management system is described. This

system includes web-based visual design tools for

constructing models and automatically generating

simulation code.

Ontological-driven simulation systems were

discussed in (Benjamin et al., 2005; Benjamin et al.,

2006). The key motivations of proposed

investigations are: to allow for the decomposition of

the target system into smaller, to distribute the model

development effort among different functional

groups and then assemble a simulation model of the

entire target system. Authors precisely discussed the

problem of simulation model interoperability

emphasizing syntactic interoperability and semantic

one. Syntactic interoperability deals with the

interoperability of implementation details (parameter

passing mechanisms, external data accesses, timing

mechanisms, for example). Semantic interoperability

“deals with the validity and usefulness of

translated/composed simulation models”. Authors in

(Benjamin et al., 2006) described a structured

ontology-driven methodology for interoperability of

simulation models.

The use of ontologies to fulfill the development of

simulation models encoding knowledge from

ontologies is presented in (Silver et al., 2007). This

paper discusses a technique that establishes

relationships between domain ontologies and a

modeling ontology and then uses the relationships to

instantiate a simulation model as ontology instances.

Techniques for translating these instances into XML

based markup languages and into executable models

for various software packages are also presented.

The use of OWL for representing object-oriented

descriptions to support distributed representations of

KEOD 2018 - 10th International Conference on Knowledge Engineering and Ontology Development

224

data, behaviors, descriptions of units and objects to be

simulated, and scenarios with initial conditions is

described in (Lacy et al., 2004). (Fishwick and Miller,

2004) provides an overview of the application of

semantic web technology to Modeling and

Simulation.

Let us consider now simulation system TriadNS

and a representation of simulation model in this

system. Later authors will discuss software tools for

simulation model development.

4 SIMULATION SYSTEM TriadNS

Simulator of computer networks TriadNS was

designed on the foundation of CAD (Computer Aided

Design) system Triad (Zamyatina et al., 2012;

Zamyatina and Mikov, 2012) in Perm State National

Research University. Software system Triad and

special language Triad (Mikov, 1995) were devoted

to the computer systems design and simulation.

The design and implementation of CAD Triad

was renewed in 2002. It was new version of Triad –

Triad.Net. New version is written in C#. Several years

later special version TriadNS for computer networks

design and analyses was implemented.

Simulation model in TriadNS is represented by

several objects functioning according to some

scenario and interacting with one another by sending

messages.

In order to build simulation model in TriadNS it

is necessary to define a structure of modeling system

(a set of objects which are connected from one to

another), the behavior of these objects (a specific

scenario). Moreover it is necessary to determine the

structure of the data exchanged between objects.

So, all objects in TriadNS may be divided into

three parts named “layers” (because we have a

hierarchy). These layers are: a layer of structures, a

layer of routines and a layer of messages

appropriately. And it is necessary to highlight that

simulation model is hierarchical one: each object

belonged to a layer of structure may be represented as

a structure of objects which belong to the lower layer

of structure and so on. The layer of routines includes

scenarios of a behavior for each object. It is important

to outline that each of objects in layer of structure

must have a scenario of behavior. A layer of structure

is convenient to present by a graph, more precisely

graph P = {U, V,W}. P-graph is named as graph with

poles. A set of nodes V presents a set of programming

objects, W – a set of connections between them, U –

a set of external poles. The internal poles are used for

information exchange within the same structure level;

in contrast, a set of external poles serves to send

messages to the objects situated on higher or

underlying levels of description.

Let us consider structure layer in TriadNS more

precisely: structure layer may be described by the

linguistic constructions of special language Triad.

Moreover, investigator may use graphical editor and

so may describe simulation model using visual

language. The example of the simulation model of

SDN (software defined) network one can see below.

Figure 1: Simulation model of SDN-network is presented

as graph, nodes of graph – computer nodes of network,

computer communication lines are presented by edges of

graph.

Simulation language Triad has several specific

features – graph constants, semantic types and etc.

Thus simulation model may be described with the

help of graph constants (with parameters): cycle(5) (5

nodes connected by edges), rectan, tree and etc.

Graph constants present the basic types of topologies

of computer network.

Besides, one may choose semantic type (Router,

Host, for example). The semantic types are used for

simulation model redefining. More details will be

given later.

An investigator may take the description of the

node’s behavior in repository (or via Internet) or may

describe it using special statements and linguistic

construction of Triad-language.

5 BASIC ONTOLOGY IN TriadNS

It is important to involve into the simulation process

not only the specialists in simulation but the specialist

in specific domains and specialists in the other

Automation of Simulation Steps using Ontological Approach

225

spheres of knowledge. That is why it is necessary to

adjust a simulation system to a specific domain.

Indeed the investigator of computer network may use

a graph theory while studying the structure of

network, or a queue network theory, or the theory of

Petri Nets. Ontologies are used in TriadNS to adjust

the simulation system to a specific domain.

Ontologies can be applied on the different steps of

simulation process (Benjamin et al., 2005; Benjamin

et al., 2006). Very often ontologies are applied for the

simulation model assembly. So the simulation model

may consist of separately designed and reusable

components. These components may be kept in

repositories or may be found via Internet. The

ontologies keep the information about

interconnections of simulation model components

and other characteristics of these components.

Moreover, ontologies enable investigators to use

one and the same terminology and etc. The basic

ontology is designed in TriadNS (fig.2.).

Figure 2: The basic ontology in TriadNS.

It’s basic classes are: TriadEntity (any named logic

entites), Model (simulation model), ModelElement (a

part of simulation model and all the specific

characteristics of a node of structure layer), Routine

(node behavior), Message (note, please, that structure

layer nodes of simulation model can interchange with

messages) and so on.

The basic properties of base ontology are:

1. A property of ownership: a model has a

structure, a structure has a node, a node has

a pole and etc.

2. A property to belong to something: a

structure belongs to the model, a node

belongs to a structure, a pole belongs to a

node and etc.

3. The properties of a pole and an arc

connection: (a) connectsWithArc (Pole,

Arc), (b) connectsWithPole (Arc,Pole).

4. A property of a node and an appropriate

routine: binding_puts On (Routine, Node).

5. The properties of a node and an appropriate

structure: (a) explicatesNode (Structure,

Node), (b) explicatedByStructure (Node,

Structure).

6. A property of a model and conditions of

simulation: binding (Model,

ModelingCondition).

The hierarchy of classes of basic ontology is

presented below (fig.3.).

6 AUTOMATION OF

SIMULATION MODEL

DEVELOPMENT

An ordinary simulation system is able to perform a

simulation run for a completely described model

only. We outlined this fact discussing the layer of

Figure 3: A hierarchy of the classes of the basic ontology in TriadNS.

KEOD 2018 - 10th International Conference on Knowledge Engineering and Ontology Development

226

structure. However, at the initial stage of designing

process an investigator may describe a model only

partly omitting description of a behaviour of one or

several nodes (routine of terminal nodes).

Moreover, simulation model may be described

without any indication on the information flows

effecting the model or without the rules of signal

transformation in the layer of messages.

However for the simulation run and the following

analysis of the model all these elements have to be

described, but this description may be inaccurate and

how inaccurate it may be we will describe below.

For example, in a completely described model

each terminal node has an elementary routine. An

elementary routine is represented by a procedure.

This procedure has to be called if one of poles of node

receives a message. But some of the terminal nodes

of partly described model do not have any routines.

Therefore the task of an automatic completion of

a simulation model consists either in “calculation” of

appropriate elementary routines for these nodes.

It was mentioned above that the routine specifies

behavioural function assigned to the node, but the

structure graph specifies additional structure level of

the model description. And at the same time, all struc-

tures must be completely described as the sub models.

These actions have to be fulfilled by the special

software tool TriadBuilder (one of the components of

simulation system TriadNS).

Software tool TriadBuilder attempts to search the

appropriate routine by the help of basic ontology (it

was described earlier). It may be found thanks to

special semantic type (semantic type “Router” and

“Host”, for example).

Model completion subsystem starts when the

internal form of simulation model is built according

to a Triad code.

First, model analyser searches the model for

incomplete nodes, and marks them. Thus, the model

analyser will mark all nodes without routines.

Then the inference module starts looking for an

appropriate routine instance for each of marked nodes

according to specification condition (the semantic

type of node and routine must coincide). An

appropriate routine may be found in repository or in

Internet.

If it is not opportunity to find appropriate routine

by a semantic type another conditions must be

checked.

It is a condition of configuration: the number of

input and output poles of node and the number of

poles of routine must coincide.

After the appropriate instance of routine has been

found, it may be put on the node.

If the condition of configuration was not met, then

the new last condition must be checked: it is a check

of the “environment” graph. An environment graph is

a graph of nodes connected to a node marked with

component TriadBuilder. Rather, the semantic type of

nodes of this graph.

7 AUTOMATION OF

SIMULATION MODEL

DEVELOPMENT FOR A

SPECIFIC DOMAIN

The simulator TriadNS has some additional special

subclasses of the basic classes (specific domain here–

computer networks):

1. ComputerNetworkModel (a model of a

computer network),

2. ComputerNetworkStructure (a structure of a

computer network model),

3. ComputerNetworkNode (a computer network

element, it contain several subclasses:

Workstation, Server, Router),

4. ComputerNetworkRoutine (a routine of a

computer and etc.

Moreover this ontology includes two special

properties of a pole.

These properties are used to check the conditions

of matching routine to a node:

1. isRequired(ComputerNetworkRoutinePole,

Boolean) – this property checks if it is

necessary to connect a pole with another

pole.

2. canConnectedWith(ComputerNetworkRout

inePole, ComputerNetworkRoutine) –this

property checks the semantic type of an

element of a structure being connected.

Nowadays SDN (software defined networks) and

SON (self - organizing networks) become popular.

These networks are the attempt to simplify and speed

up the planning, configuration, management,

optimization of communications networks.

Automation of Simulation Steps using Ontological Approach

227

Figure 4: A hierarchy of the classes of the ontology

including classes and subclasses describing computer

networks in TriadNS.

The investigation of the routing algorithm

SBARC for SDN were carried out.

Thus it was necessary to design several new

subclasses:

1. subclass (SDNNode) of class

ComputerNetworkNode,

2. subclass SBARCRoutine,

3. subclass SDNNodeRoutine of class

ComputerNetworkRoutine.

One can see a hierarchy of the ontology classes

describing computer networks in TriadNS, in fig.4.

So one can see that including new classes and

subclasses to basic ontology allows to expand the

possibilities of the simulation system. Moreover, the

model is validated and it is carried out automatically.

Let us consider another example: transformation

of conceptual simulation model to a model described

with the help of another visual language.

8 SIMULATION MODEL

TRANSFORMATION

Usually simulation systems allow investigating

objects (or situation) with the help of mathematical

scheme – queuing networks. But it is advisable to

apply other mathematical schemes, for example, Petri

nets or Markov processes.

Special software components of TriadNS

simulation system (embedded software tools) allow

transforming conceptual model to the model which is

described with special visual languages (MDE

(model driven engineering approach) (Chetinskaya

and Verbraeck, 2011) implemented in TRIADNS.

One can see the model of simple computer network

consists of two workstations and one server and this

transformed model. This model presents as Petri Net.

Figure 5: Conceptual model of computer network is

transformed to Petri net.

In order to fulfill model transformation the basic

ontology was enriched with new classes and

subclasses, and the TriadNS software tools – with

software that implemented the rules for translating

one model to another. The ontology for Petri Net is

presented below.

KEOD 2018 - 10th International Conference on Knowledge Engineering and Ontology Development

228

Figure 6: Conceptual model of computer network is

transformed to Petri net.

9 CONCLUSIONS

So we considered ontological approach using by the

developers of simulation system.

Ontological approach was applied for automation

of several steps of simulation process: redefining of

simulation model, adjusting of conceptual simulation

model to a specific domain (embedded DSL

component). Moreover one more very important step

of simulation process was implemented in TriadNS:

verification and validation. Special software

component build ontology of errors. After the step of

translation this component carries out the elimination

of errors due to rules specified in ontology.

Thus the automation of these steps allows

investigators to carry out simulation in convenient

software environment, to involve in simulation

process specialists in various specific domains and in

this way to receive more adequate simulation models

and appropriate results. Moreover ontological

approach help to reduce the overall time needed for

simulation.

ACKNOWLEDGMENTS

The reported study was funded by RFBR according

to the research project № 18-01-00359.

REFERENCES

Balci, O., 1998. Verification, Validation, and

Accreditation. Proceedings of the 1998 Winter

Simulation Conference. D.J. Medeiros, E.F. Watson,

J.S. Carson and M.S. Manivannan, eds., Grand Hyatt

Washington, Washington DC, pp.41-48.

Balci, O., Nance, R. E., Arthur J. D., 2002. Expanding Our

Horizons In Verification, Validation And Accreditation

Research and Practice Proceedings of the 2002 Winter

Simulation Conference E. Yücesan, C.-H. Chen, J. L.

Snowdon, and J. M. Charnes, eds., WSC 2002, San

Diego, California, pp.653-663.

Law, A. M., McComas, M. G., 2001. How to build valid

and credible simulation models. In Proc. 2001 Winter

Simulation Conf., ed. B.A. Peters, J. S. Smith, D. J

Medeiros, and M. W. Rohrer, Piscataway, New Jersey:

IEEE. pp. 22-29.

Sargent, R. G., 2005. Verification and validation of

simulation models. Proceedings of Winter Simulation

Conf., ed. M. E. Kuhl, N. M. Steiger, F. B. Armstrong,

and J. A. Joines, Piscataway, New Jersey, 2005: IEEE

pp.130-143.

Sargent, R. G. 2007. Verification and Validation of

Simulation Models. Proceedings of Winter Simulation

Conf., S. G. Henderson, B. Biller, M.-H. Hsieh, J.

Shortle, J. D. Tew, and R. R. Barton, eds., J.W. Marriott

Hotel, Washington, 2007, pp.124-137

Salmon, S., Aarag, H., 2011. Simulation Based

Experiments Using Ednas: The Event-Driven Network

Аrchitecture Simulator. In Proceedings of the 2011

Winter Simulation Conference S. Jain, R.R. Creasey, J.

Himmelspach, K.P. White, and M. Fu, eds.. The 2011

Winter Simulation Conference 11-14 December 2011

Grand Arizona Resort Phoenix, AZ, pp. 3266-3277

Cetinkaya, D., Verbraeck A., 2011. Metamodeling and

model transformations in modeling and simulation //

Proceedings of the 2011 Winter Simulation Conference,

2011, pp. 3048-3058

Sokolov, B. V., Uysupov, R. M. 2005. Konceptualnye

osnovy kvalimetrii modeley i polimodelnyh

kompleksov//Imitacionnoe modelirovanie. Teoria I

praktika. Sbornik dokladov vtoroi vserossiyskoy

nauchno-prakticheskoi konferencii IMMOD-2005.

Tom 1. Spb..: CNIITS. 2005.-s.65-70.

Zamyatina, E. B., Mikov, A. I., Mikheev, R. A., 2012.

Lingvisticheskie i intellektualnye instrumentalnye

sredstva simulyatora kompiutornyh setei TRIADNS.

International Journal “Information theories &

Applications (IJ ITA). Vol 19, Number 4, 2012, pp.355-

368. ITHEA, Sofia, 1000, P.O.B. 775, Bulgaria.

Zamyatina, E. B., Mikov, A. I., 2012. Programmnye

sredstva imitatsii Triad.Net dlya obespecheniya eyo

Automation of Simulation Steps using Ontological Approach

229

adaptiruemosti I otkrytosti. Informatizatsiya i svyaz.

№5, 2012, ANO «Redaktsiya zhurnala

«Informatizatsiya I svyaz», ISSN 2078-8320, PP.130-

133[in Russian].

Mikov, A. I., 1995. Formal Method for Design of Dynamic

Objects and Its Implementation in CAD Systems //

Gero J. S. and F. Sudweeks F.(eds), Advances in

Formal Design Methods for CAD, Preprints of the IFIP

WG 5.2 Workshop on Formal Design Methods for

Computer-Aided Design, Mexico, Mexico, 1995.

pp.105-127.

Benjamin, P., Akella, K. V., Malek, K., Fernandes, R. 2005.

An Ontology-Driven Framework for Process-Oriented

Applications // Proceedings of the 2005 Winter

Simulation Conference / M. E. Kuhl, N. M. Steiger, F.

B. Armstrong, and J. A. Joines, eds.,– pp 2355-2363

Benjamin, P., Patki, M., Mayer, R. G. 2006. Using

Ontologies For Simulation Modeling // Proceedings of

the 2006 Winter Simulation Conference/ L. F. Perrone,

F. P. Wieland, J. Liu, B. G. Lawson, D. M. Nicol, and

R. M. Fujimoto, eds. –pp.1161-1167

Mikov, A., Zamiatina, E., 2014. Program Tools and

Language for Networks Simulation and Analysis.

Proceedings SDN & NFV – The Next Generation of

Computational Infrastructure: 2014 International

Science and Technology Conference «Modern

Networking Technologies (MoNeTec)» October 27-29,

2014 Lomonosov Moscow State University, pp. 94-102

Mikov, A., Zamyatina, E., Kubrak, E., 2007.

Implementation of simulation process under incomplete

knowledge using domain ontology. In proceedings of 6-

th EUROSIM Congress on modeling and Simulation. 9-

14, September, 2007, Ljubljana, Slovenia, Vol.2. Full

papers, 7 p.

Sarli, J. L., 2005. Ontology-Based Modeling Framework to

Generate Federation Object Model in the Supply Chain

Domain. Proceedings of Winter Simulation Conf., ed.

W. K. V. Chan, A. D'Ambrogio, G. Zacharewicz, N.

Mustafee, G. Wainer, and E. Page, eds., Piscataway,

New Jersey, 2005: IEEE pp.4636-4637.

Beck, H., Morgan, K., Jung, Y., Grunwald, S., Kwonb, H.,

Wub, J., 2010. Ontology-based simulation in

agricultural systems modeling. Agricultural Systems,

103 (2010) pp.463–477.

Zaletelja, V., Vrabičb, R., Hozdićb, E., Butalab, P. A.,

2018. foundational ontology for the modelling of

manufacturing systems. Advanced Engineering

Informatics, 38, 2018, pp. 129–141

Lacy, L., Gerber, W., 2004. Potential Modeling and

Simulation Applications of the Web Ontology

Language – OWL. In Proceedings of the 2004 Winter

Simulation Conference, eds. R. G. Ingalls, M. D.

Rossetti, J. S. Smith, and B. A. Peters, pp.265–270.

Silver, G. A., O. A. Hassan, and J. A. Miller. 2007. From

Domain Ontologies to Modeling Ontologies to

Executable Simulation Models. In Proceedings of the

2007 Winter Simulation Conference, eds. S. G.

Henderson, B. Biller, M. H. Hsieh, J. Shortle, J. D. Tew,

and R. R. Barton, pp. 1108–1117.

Fishwick, P. A. and J. A. Miller. 2004. Ontologies for

Modeling and Simulation: Issues and Approaches. In

Proceedings of the 2004 Winter Simulation Conference,

eds. R. G. Ingalls, M. D. Rossetti, J. S. Smith, and B. A.

Peters, pp. 259–264.

KEOD 2018 - 10th International Conference on Knowledge Engineering and Ontology Development

230