2015 10th System of Systems Engineering Conference (SoSE)

Simulating SysML Models: Overview and Challenges

Mara Nikolaidou, George-Dimitrios Kapos, Anargyros Tsadimas,
Vassilios Dalakas and Dimosthenis Anagnostopoulos
Department of Informatics and Telematics
Harokopio University of Athens
Athens, GREECE
{mara, gdkapos, tsadimas, vdalakas, dimosthe}@hua.gr

system models for the specific domain, based on the profile

Abstract – SysML language, proposed by OMG, is a
specifications.
commonly accepted standard to model and study systems-
of-systems (SoS). It provides the means to depict SoS
As simulation is a common method for estimating the
components and their behavior in a multi-layer fashion
performance of systems, there is currently strong interest in
and explore alterative architectures for their design. To
generating simulation code from SysML models. Recent
validate SysML models in terms of performance criteria,
efforts (as for example [7], [8], [9]) provide the ability to
simulation is usually the preferred method employed. To
generate executable simulation code for different
this end, different SysML diagrams are utilized, while
simulation languages or environments (as for example
numerous simulation methodologies and tools are
Arena, Modelica or DEVS). In most of these efforts,
employed. There are many efforts targeting simulation
profiling mechanism is used to embed simulation properties
code generation from SysML models. Model-based system
into SysML models. Simulation-specific profiles are
engineering concepts are adopted in most of them to
employed in a popular modeling tool, such as Magic Draw
generate simulation models from SysML models.
or IBM Rational Modeler mentioned above, to annotate
Nevertheless, this process is not standardized, although
SysML models with simulation properties appropriate for
most of current approaches tend to follow the same steps,
the specific simulator. Afterwards, the corresponding
even if they employ different tools. The scope of this paper
enriched SysML models are transformed to executable
is to provide a comprehensive understanding of the
simulation code for the specific environment. Model
similarities and differences of existing approaches and
transformation languages, such as ATL [10] and QVT [11],
identify current challenges in fully automating SysML
are often utilized to transform SysML models to simulation
models simulation process.
models represented in XMI, an XML representation
language for UML/SysML models. In such cases, model-
Keywords: SysML, simulation, automated code
based engineering techniques are adopted and model-
generation, model transformation, model-based system
driven software generation is employed according to MDA
engineering.
standards [12]. The benefit of employing such standards to
1 Introduction accomplish simulation code generation is the fact that all
the steps leading to the creation of executable simulation
SysML ([1]) is a language commonly used for model-based models are independent of the tools used for SoS modeling
system design (MBSD). It is an Object Management Group and engineering and the specific system domain and the
(OMG) standard that supports specification, analysis, methodology employed and software created are only
design, verification and validation of a broad range of restricted by the simulation methodology/environment.
systems and systems-of-systems. It provides discrete
diagrams to describe system structure and components, to Although the process of generating simulation code is
explore allocation policies crucial for system design and to similar in the aforementioned approaches, it is not
identify design requirements. It is widely applied for standardized. Furthermore, there are still restrictions in
systems-of-systems (SoS) engineering [2]. fully automating the simulation code generation process
related to the system domain, the existence of pre-defined
SysML is a general-purpose language, facilitating modeling simulation libraries of for specific domain and the
of any system or system-of-systems. UML profiles can be characteristics of the simulation methodology/environment.
employed to restrict or extend SysML features to serve a To this end, a thorough overview of different approaches
specific domain, as for example real-time and embedded for simulating SysML models is conducted in this paper, in
systems [3] or information systems [4]. These profiles, order to identify similarities and basic differences, in a
accompanied with specific plugins, can be executed within effort to standardize the process of simulating SysML
UML modeling tools (such as Magic Draw [5] or IBM models taking into account that different tools and methods
Rational Modeler [6]) and are capable of producing valid

Authorized licensed use limited to: Otto-von-Guericke Universitaet Magdeburg. Downloaded on June 10,2024 at 15:15:19 UTC from IEEE Xplore. Restrictions apply.
978-1-4799-7611-9/15/$31.00 ©2015 IEEE 328
 2015 10th System of Systems Engineering Conference (SoSE)

are to by used. The approaches studied were selected based Focusing on the description of specific domain entities,
on two criteria: specialized SysML profiles are introduced, while the need
to manage performance requirements during system design
a) They are based on model-based system engineering is also addressed. SysML provides the means for
concepts and apply MDA standards as discussed above to requirement description, while there are efforts, such as [4],
ensure compatibility with current SoS engineering [15], [17] focusing on requirement verification described
standards. Manual or semi-automated generation of using SysML.
executable simulation models from SysML models can be
cumbersome, tedious and error-prone. In addition, building In any case, to validate SysML models in terms of
custom tools for this purpose restricts reusability and performance, they should be simulated first. Apparently
interoperability with other simulation platforms. SysML supports a variety of diagrams describing system
structure and states, necessary to perform simulation, which
b) Different simulation methodologies/environments are are utilized by different approaches ([18], [19]). In most
utilized to ensure the generality of the deduced conclusions. cases, SysML models defined within a modeling tool are
The simulation techniques selected may serve different exported in XMI format and, consequently, transformed
system domains, implement either discrete or continuous into simulator specific models to be forwarded to the
simulation and utilize different SysML diagrams to simulation environment. To embed simulation-specific
integrate simulation characteristics into SysML models. properties within SysML models, profiles are introduced.
In all cases, stereotypes and constraints defined within the
Besides the simulation code generation process, the system profile are related to the simulation platform employed ([7],
validation process after the simulation is completed is also [8], [9], [15], [18], [19]). Simulation model validity may be
explored. In this case, simulation data are used to verify ensured by applying constraints in the models produced by
performance requirement defined during system design. the profile using declarative languages, as OCL ([7], [8],
There are approaches, such as [13], [14], [15], providing [3], [20]) or even java plugins ([14]).
comprehensive solutions in assisting the system engineer
identifying system design conflicts or drawbacks. Some of them are general ([9], [18]) constraint only by the
However, the automated and transparent integration of the simulation platform and facilitating simulation of any kind
system validation process within the SysML model used of of systems. In such case both the system structure and the
system design with the modeling tool remains a challenge. system behavior described in terms of the simulation
environment must by defined. Both discrete event (for
The rest of the paper is organized as follows: In Section 2 example [9]) and continuous (for example [13]) simulation
related work is discussed. In Section 3 existing approaches environments are supported, utilizing different SysML
for simulating SysML models are discussed, while their diagrams, e.g activity and state diagrams in the first case
basic characteristics are explored in a comparative study. and parametric diagrams in the second case to model
Challenges in automated simulation code generation and system behavior. In the case of general approaches,
system validation are identified in Section 4. simulation code generation is usually not fully automated
in terms of system behavior, which is restricted in the
profile in term of functionality and expressiveness.
2 Related work
There are many efforts that employ SysML for model- Most approaches are focused on a specific system domain,
based system design in different domains. Recently, SLIM while profiles contain stereotypes to describe specific
[16], a commercial collaborative model-based systems domain components, while their behavior is prescribed in
engineering workspace that uses SysML as the front-end simulation libraries contained in the simulation
for orchestrating system engineering activities from the environments selected. Popular examples of such systems
early stages of system development, is available from are presented for example in [17] for embedded systems
Intercax. The SysML-based system model serves as a simulated using Modelica or in [7] for manufacturing
unified, conceptual abstraction of the system, independent assembly systems simulated using Arena. Depending on
of the specific design and analysis tools that shall be used the nature and specific characteristics of the specific
in the development process. It is designed to provide domain, there is a diversity of ways proposed to simulate
plugins to integrate the system model to a variety of design SysML models, utilizing different diagrams.
and analysis tools. Until now, only the integration of
SysML and other model repositories, such as product Existing approaches may also be grouped in an alternative
lifecycle management (PLM) tools is implemented. fashion, depending on whether or not they are utilizing
Integration with MATLAB/Simulink, Mathematica and current model-driven software engineering standards for
OpenModelica is offered in a variety of commercial tools, simulation code generation. Custom tools, not utilizing
but automated simulation code generation is not existing standards, although flexible and fast, do not
implemented yet. promote model transformation validation, while they
restrict reusability and interoperability with other

Authorized licensed use limited to: Otto-von-Guericke Universitaet Magdeburg. Downloaded on June 10,2024 at 15:15:19 UTC from IEEE Xplore. Restrictions apply.
329
 2015 10th System of Systems Engineering Conference (SoSE)

simulation platforms. Many approaches are grouped in this MATLAB Stateflow simulation environment. No detailed
category, such as those presented in [3], [4], [20]. information on MOF 2.0 based meta-models for all these
environments are provided, though MDA principles are
In order to follow model-driven code generation principles adopted by the authors. In [27], the MDEReqTraceTool,
and facilitate SysML-to-Simulation model transformation, currently under development, is proposed to integrate
a meta-model describing simulation entities must be requirement verification information, obtained using
defined. The existence of such a simulation meta-model is external tools, within SysML system models by updating
imperative to facilitate the transformation of SysML corresponding SysML requirement verification matrixes.
models described in XMI format into simulation models Such a feature will enable the MARTE requirements
[21]. To ensure compatibility with UML/SysML related verification using external tools, in a transparent fashion for
standards, MOF 2.0, the meta-model defined by OMG to the system designer working with MARTE/SysML models.
define them, should be also used for simulation meta-model
description. MOF 2.0 compliance enables the application of CASSI Tool
standard languages, such as ATL or QVT to program
In [4] SysML extensions were proposed for information
SysML-to-Simulation model transformation. Providing a
system design, which are implemented within the context
MOF 2.0 meta-model in XMI format for a simulation
of a custom, in-house tool called CASSI. CASSI targets
methodology or tool, such as those defined for Modelica
information system integration, while three different design
[22] or DEVS [23], enhances the transformation process
views are supported, depicted using SysML external and
and facilitates the usage of the specific simulation methods
internal block diagrams. The behavior of system
for SysML model simulation. The transformation needs to
components is described within CASSI using sequence
be defined only once for a pair of domain and simulation
diagrams, transformed to a simulation model based on
environment.
Petri-nets, which is executed by an external simulator.
Although based on MDA principles, existing standards and
3 A Comparative Overview of SysML tools are not utilized. CASSI is based entirely on custom
Model Simulation Approaches tools.

Out of the extensive literature for SysML model As described on [28], information system configurations
simulation, the approaches presented in the following were defined using CASSI are evaluated using simulation to
chosen taking into account the following criteria: a) they verify performance and availability requirements. NFR
are based on model-driven software engineering, b) they verification is performed by the system designer using
support all the steps discussed, i.e. profile definition, external tools by the aid of Service Level Objective (SOL)
automated simulation code generation and system concept, while system validation and requirement
validation and c) they utilize different simulation verification results are not integrated within SysML system
environments. model.

MARTE Profile and Related Tools TTool Toolkit and Related Efforts

MARTE is a UML profile proposed by the OMG [24] in TTool Toolkit (http://ttool.telecom-paristech.fr) integrates
2009 to support model-based design of real-time and numerous tools targeting real-time embedded system
embedded systems. It focuses on the performance and engineering. AVATAR SysML profile is one of them,
scheduling properties of real-time systems. Performance targeting safety and security properties of embedded
and scheduling requirements are modeled as constraints systems [31]. TEPE, a graphical expression language based
defined using VSL, a language for formulating on SysML parametric diagrams, is introduced for
semantically well-formed algebraic and time expressions. representing requirements making them amenable to
automated verification [29]. The profile also enables the
After SysML standardization, there are numerous efforts to definition of system behavior through state machine
combine SysML and MARTE profiles (for example [3], diagrams. Model verification is performed using a
[25], [26], [27]). Basically they focus on integrating SysML constraint language called UPPAAL (based on OCL), to
requirements and VSL language, employed to specify ensure system model validity before simulating them [30].
them. VSL well-defined semantics enable the automated DIPLODOCUS, a simulation engine targeting on System-
verification of corresponding SysML requirements using on-Chip design, is integrated in TTool. It is based on Y-
external tools. Chart simulation approach, based on timed-automata. The
IFx toolkit3 [32] also provides simulation capabilities
In [26] and [27] the effort presented is focuses on within TTool framework.
generating executable code in SystemC, a language for
describing executable software for embedded systems Model-driven engineering concepts are introduced in
using model transformation techniques. Furthermore, the TTool toolkit components to automatically generate
same methodology is suggested to provide executable simulation code based on predefined libraries for the
models for Promela/SPIN model checking environment and domain of embedded systems. Though, all the tools

Authorized licensed use limited to: Otto-von-Guericke Universitaet Magdeburg. Downloaded on June 10,2024 at 15:15:19 UTC from IEEE Xplore. Restrictions apply.
330
 2015 10th System of Systems Engineering Conference (SoSE)

developed are proprietary to work within TTool DevSys Framework

environment, while no MOF 2.0 compatible meta-model is
defined for the simulation or model checking The authors have proposed DEVSys framework for the
environments. Requirement verification is facilitated by simulation of SysML model using DEVS [9]. A DEVS
external tools, such as IFx toolkit3. SysML profile is defined for the enrichment of SysML
models with all required properties to generate the classical
SysML-to-Arena Transformation Tools DEVS simulation models [34]. Block definition and
internal block diagrams are utilized to describe system
In [7], manufacturing line system models are defined in structure, while state machine, activity and parametric
SysML and transformed using ATL to be simulated using diagrams are utilized to define system behavior for
Arena simulation software. With the definition of a SysML simulation purposes. To this end, a MOF 2.0 meta-model
profile, Arena-specific properties modeling manufacturing for DEVS is proposed and applied for the definition of a
systems are incorporated within SysML block definition standards-based QVT transformation of enriched SysML
and activity diagrams [33]. Corresponding ARENA models to DEVS models that are consequently transformed
simulation libraries are incorporated with ARENA to executable DEVS code [35]. The profile is not restricted
environment, and properly instantiated to construct the to a specific system domain, enabling the simulation of any
simulation model executed within ARENA tool. As far as system described in SysML following DEVS behavioral
simulation is concerned only system structure is defined in model. Thus, the main restriction of this approach is that
SysML diagrams. System simulation behavior is defined the system designer should be aware of DEVS
within ARENA manufacturing system libraries. SysML-to- methodology and concepts to properly define system
ARENA model transformation is performed using ATL behavior. Constraints defined in OCL and Java plugins are
based on model-based software engineering principles, available in the profile to ensure model validity before
while a corresponding MOF-based meta-model for simulation.
ARENA manufacturing system libraries is defined. The
exploitation of simulation output towards system model The combination of DEVSys framework with EIS profile
validation is not discussed. for information system design is presented in [14]. In this
case, information system component libraries were
SysML4Modelica Project implemented within DEVS, while SysML-to-DEVS QVT
transformation was utilized only to generate the simulation
The SysML4Modelica profile endorsed by OMG [8]
code corresponding to system structure. System behavior
enables the transformation of SysML models to executable
was already implemented in DEVS libraries. Additionally,
Modelica simulation code. To embed simulation
performance-related attributes defined of all system
capabilities within SysML, ModelicaML profile is used
component are calculated during simulation and embedded
[22]. QVT is used for the transformation of SysML models
within the SysML model after the completion of the
defined using ModelicaML profile to executable Modelica
simulation.
models. A corresponding MOF 2.0 meta-model for
Modelica is defined. The overall approach is fully Summary
compatible with model-driven engineering concepts,
making it suitable of efficient SoS engineering. The basic features of the reviewed approaches are
summarized in Table 1. A variety of simulators are
In [17] focus is given on using the SysML4Modelica utilized, selected usually based on the system domain and
profile for embedded systems engineering. In the proposed their popularity. Custom solutions tend to be avoided. None
profile, SysML requirement entity is extended with testable of the presented works fully covers all features effectively.
characteristics. Testable requirements are associated to However, the adaptation of model-based engineering
conditions under which the requirement is verified with the principles and the utilization of standard language to
use of experiments or test cases. Verification conditions are transform SysML-to-Simulation model is clearly gaining
defined as part of a test case, which in turn may be momentum. Most of the approach, even if defined to be
simulated using Modelica simulation language in external general, focus on a specific system domain and provide
simulators to ensure that a design alternative satisfies corresponding model libraries within the simulation
related requirements [13]. Requirement verification is environment to simply the transformation process. In most
performed in an external modelica tool (MathModelica) case model simulation is closely related to corresponding
through visual diagrams created during simulation. requirement verification, while most of the approaches
tackle with requirement verification as well, if though in
The only possible limitation of this approach is that it is
many case this process in performed outside the SysML
tailor-made to the Modelica in order to describe the
modeling tool. Finally, the automated simulation code
requirements and the verification process together.
generation capabilities offered are constantly increasing.

Authorized licensed use limited to: Otto-von-Guericke Universitaet Magdeburg. Downloaded on June 10,2024 at 15:15:19 UTC from IEEE Xplore. Restrictions apply.
331
 2015 10th System of Systems Engineering Conference (SoSE)

Table 1: A Comparative Overview of SysML Simulation Approaches

Profile Features Offered
System Simulator Profile Characteristics Transforma MDA Code SysML model validation/
domain tion conforma generation requirement verification
language nce support
Requirement verification
Real-Time - Focus on performance and time performed within SysML models
MARTE Non- Non Non
Embedded requirement description and Medium (MDEReqTrace integrates
Profile Specified Specified specified
Systems verification utilizing VSL requirement verification data from
external tools within SysML)
- Focus on describing performance
Requirement verification
Information requirements Non Semi
CASSI Tool Systems
Petri-Nets
- System behavior is described using Specified
Low
automated
performed by an custom external
tool
Sequence diagrams
- Focus on requirement description using
Real-Time Y-Chart,
TTool TEPE Non Fully Requirement verification
Embedded Timed- Medium
Toolkit - System behavior is described using Specified automated performed by external tools
Systems Automata
State Machine diagrams
Manufacturi - Focus on the description of the specific
SysML to Fully
ng Line Arena domain to incorporate simulation- ATL High Not Specified
Arena Tools Systems automated
related characteristics
- Focus on describing performance
General
requirements
SysML4 (emphasis Fully Requirement verification
Modelica - System behavior under exploitation is QVT High
Modelica on real-time automated performed by external tools
defined as Test Cases using Modelica
systems)
ML
- Focus on embedding simulation output
General
within SysML models
DEVSys (case study: Fully Requirement verification
DEVS - System behavior is described using QVT High
Framework Information automated performed within SysML models
State Machine, Parametric and
Systems)
Activity diagrams

4 Challenges Identified initiatives, as those employed by Modelica and DEVS

community are already successful.
Having in mind existing approaches, it is evident that there
is a strong interest in simulating SysML models in an c) Utilize simulation output to validate SysML models and
automated fashion to serve SoS engineering and especially verify corresponding requirements defined in such
SoS design. Since different system domains should be models. In order to simplify requirement verification
effectively supported, it is expected that different process, we endorse the suggestion of SLIM to conduct
simulation methods and tools will be employed. Though, it requirement verification within SysML modeling tools,
is imperative that a standardized methodology/framework, independently of the simulation methods and tools. The
based on OMG standards, should be proposed to guide incorporation of simulation results within the SysML
experts to develop tools targeting specific domains and model should be facilitated for this purpose. Such
simulation environments. Most recent approaches seam to enhancements simplify the evaluation process, allowing
follow the same basic steps: the system designer to focus on the examination of the
unverified requirements and, consequently, the
a) Define simulation/domain specific profiles. In this detection of the necessary solution re-adjustments.
process, efforts should concentrate on define simulator-
specific profiles that may be combined with domain As derived from the examination of existing approaches,
specific profiles. Furthermore, the exploration of a there are two key issues in requirements verification during
simulator-agnostic profile is suggested for discrete- model-based system design that have not been fully
event and continuous simulators respectively, taking addressed: (a) the estimation of system models behavior in
into account that existing approaches utilize the same a generic and -at the same time- automated manner, and (b)
SysML diagrams. the designation of the requirements that have not been
verified in the original system model. Precisely, although
b) Transform SysML to simulation models in a simulation has been identified as an appropriate technique
standardized fashion, utilizing languages as QVT and for the estimation of system models’ performance, obtained
simulation results should be incorporated within the
ATL. Simulator-specific profiles should be
original system model and comparison against the
accompanied by corresponding MOF-based meta-
predefined, performance-related, requirements should be
models for the corresponding simulators. The definition
of such meta-models openly available may also performed within the modeling environment.
promote simulator interoperability. Corresponding

Authorized licensed use limited to: Otto-von-Guericke Universitaet Magdeburg. Downloaded on June 10,2024 at 15:15:19 UTC from IEEE Xplore. Restrictions apply.
332
 2015 10th System of Systems Engineering Conference (SoSE)

[19] O. Schonherr, O. Rose, First steps towards a general SysML

References model for discrete processes in production systems, in:
[1] OMG, Systems Modeling Language (SYSML) Proceedings of the 2009 Winter Simulation Conference, Austin,
Specification, Version 1.3 (June 2012). URL TE, USA, 2009, pp. 1711–1718.
http://www.omg.org/spec/SysML/1.3/PDF [20] S. Chouali, A. Hammad, H. Mountassir, Assembling
[2] INCOSE, Systems Engineering Handbook, version 3.2.2 components using sysml with non-functional requirements,
Edition, International Council on Systems Engineering, San Electronic Notes in Theoretical Computer Science 295 (0) (2013)
Diego, CA, USA, 2012. 31 – 47, Proceedings the 9th International Workshop FESCA.
[3] Quadri, I.R., Sadovykh, A, Indrusiak, L.S, "MADES: A [21] M. Nikolaidou, G.-D. Kapos, V. Dalakas, D.
Mixed SysML/MARTE methodology for real-time and embedded Anagnostopoulos, Basic Guidelines for Simulating SysML
avionics systems," ERTS 2012 Conference in Toulouse, France, Models: An Experience Report, in: Proc. Seventh Int. Conf. on
Feb. 2012. System of Systems Engineering (SoSE) 2012, 2012, pp. 95–100.
[4] S. Izukura, K. Yanoo, T. Osaki, H. Sakaki, D. Kimura, J. [22] W. Schamai, Modelica Modeling Language (ModelicaML):
Xiang, Applying a model-based approach to IT systems A UML Profile for Modelica, Tech. rep. (2009).
development using SysML extension, in: MoDELS, Vol. 6981 of [23] S. Mittal and J. L. Risco Martín, Netcentric System of
Lecture Notes in Computer Science, Springer, 2011, pp. 563–577. Systems Engineering with DEVS Unified Process, CRC Press,
URL http://dx.doi.org/10.1007/978-3-642-24485-8 2013.
[5] NoMagic, SysML Plugin for Magic Draw. URL [24] OMG, UML profile for MARTE: Modeling and analysis of
http://www.nomagic.com/products/magicdraw-addons/ sysml- real-time embedded systems specification, version 1.0, Nov 2009.
plugin.html
[25] H. Espinoza, D. Cancila, B. Selic, S. Gerard, Challenges in
[6] IBM, Rational Software Modeler. URL combining SysML and MARTE for model-based design of
http://www.ibm.com/developerworks/rational embedded systems, in: ECMDA-FA, Vol. 5562 of Lecture Notes
[7] O. Batarseh, L. F. McGinnis, System modeling in SysML in Computer Science, Springer, 2009, pp. 98–113.
and system analysis in ARENA, in: Proceedings of the Winter [26] Rota Sena Marques, M.; Siegert, E.; Brisolara, L.,
Simulation Conference, WSC ’12, 2012, pp. 258:1–258:12. "Integrating UML, MARTE and sysml to improve requirements
[8] OMG, SysML-Modelica Transformation (SyM) (Nov. specification and traceability in the embedded domain," Industrial
2012).URL http://www.omg.org/spec/SyM/1.0/PDF/ Informatics (INDIN), 2014 12th IEEE International Conference
[9] G.-D. Kapos, V. Dalakas, M. Nikolaidou, D. on, pp.176,181, 27-30 July 2014.
Anagnostopoulos, An integrated framework for automated [27] Marcello Mura, Amrit Panda, Mauro Prevostini,
simulation of SysML models using DEVS, Simulation 90 (6) “Executable Models and Verification from MARTE and SysML:
(2014) 717–744. a Comparative Study of Code Generation Capabilities”. MARTE
[10] Atlas Transformation Language. URL https://eclipse.org/atl/ Workshop (2008).
[11] OMG, MOF 2.0 Query/View/Transformation Language ver. [28] D. Kimura, T. Osaki, K. Yanoo, S. Izukura, H. Sakaki, A.
1.1, Jan. 2011. URL http://www.omg.org/spec/QVT/1.1 Kobayashi, “Evaluation of IT systems considering characteristics
as system of systems”, 6th International Conference o System of
[12] OMG, Model-Driven Architecture. URL Systems Engineering (SoSE), 2011.
http://www.omg.org/mda/
[29] D. Knorreck, L. Apvrille, P. de Saqui-Sannes, TEPE: A
[13] W. Schamai, P. Helle, P. Fritzson, C. J. J. Paredis, Virtual sysml language for time-constrained property modeling and
verification of system designs against system requirements, in: formal verification, SIGSOFT Softw. Eng. Notes 36 (2011) 1–8.
Proceedings of the 2010 international conference on Models in
software engineering, MODELS’10, Springer-Verlag, Berlin, [30] G. Behrmann, A. David, K. Larsen, J. Hakansson, P.
Heidelberg, 2011, pp. 75–89. Petterson, W. Yi, M. Hendriks, UPPAAL 4.0, in: Quantitative
Evaluation of Systems, 2006. QEST 2006. Third International
[14] A Tsadimas, GD Kapos, V Dalakas, M Nikolaidou, D Conference on, 2006, pp. 125–126.
Anagnostopoulos, "Integrating Simulation Capabilities into
SysML for Enterprise Information System Design", IEEE 8th [31] G. Pedroza, L. Apvrille, D. Knorreck, Avatar: A SysML
Conference on SoSE, Adelaide, Australia, June 2014. environment for the formal verification of safety and security
properties, in: NOTERE 2011, 11th Annual International
[15] J.-F. Petin, D. Evrot, G. Morel, P. Lamy, Combining Conference on, 2011, pp. 1–10.
SysML and formal methods for safety requirements verification,
in: 22nd International Conference on Software & Systems [32] I. Ober, S. Graf, I. Ober, Validating timed UML models by
Engineering and their Applications, Paris, France, 2010. simulation and verification, International Journal on Software
Tools for Technology Transfer 8 (2) (2006) 128–145.
[16] M. Bajaj, D. Zwemer, R. Peak, A. Phung, A. Scott, M.
Wilson, Slim: collaborative model-based systems engineering [33] L. F. McGinnis, E. Huang, K.S. Kwon, V. Ustun,
workspace for next-generation complex systems, in: Aerospace “Ontologies and simulation: a practical approach”, Journal of
Conference, 2011 IEEE, 2011, pp. 1–15. Simulation, 08/2011; 5:190-201.
[17] A. A. Kerzhner, J. M. Jobe, C. J. J. Paredis, A formal [34] B. P. Zeigler, H. S. Sarjoughian, Introduction to DEVS
framework for capturing knowledge to transform structural Modeling and Simulation with JAVA (2003).
models into analysis models, Journal of Simulation 5 (3) (2011) [35] G.-D. Kapos, V. Dalakas, A. Tsadimas, M. Nikolaidou, D.
202–216. Anagnostopoulos, Model-based system engineering using SysML:
[18] L. McGinnis, V. Ustun, A simple example of SysML-driven Deriving executable simulation models with QVT, in: Systems
simulation, in: Winter Simulation Conference (WSC), Conference (SysCon), 2014 8th Annual IEEE, IEEE, 2014, pp.
Proceedings of the 2009, IEEE, 2009, pp. 1703–1710. 531–538.

Authorized licensed use limited to: Otto-von-Guericke Universitaet Magdeburg. Downloaded on June 10,2024 at 15:15:19 UTC from IEEE Xplore. Restrictions apply.
333