Development of Ontologies for Reasoning and Communication in
Multi-Agent Systems
Sebastian Törsleff, Constantin Hildebrandt and Alexander Fay
Institute of Automation Technology, Helmut-Schmidt University, Hamburg, Germany
Keywords: Ontology Development, Ontology Engineering, Model-driven Engineering, Multi-Agent Systems,
Cyber-physical Systems.
Abstract: In future cyber-physical systems, such as smart factories and energy grids, ontologies can serve as the enabler
for semantically precise communication as well as for knowledge representation and reasoning. Multi-agent
systems have shown to be a suitable software development paradigm for cyber-physical systems and may
well profit from harnessing ontologies in terms of reduced engineering effort and better interoperability. This
contribution presents a development methodology for ontologies that enable communication and reasoning in
Multi-Agent Systems for cyber-physical systems. The methodology is unique in addressing a set of
requirements specific to this application domain.
1 INTRODUCTION
Multi-agent systems (MASs) have the potential to
play a key role in enabling future cyber-physical
systems for smart factories and energy grids (Leitao
et al., 2016; Vrba et al., 2014). Major benefits of
MASs are modularity, flexibility and robustness.
Ontologies can provide value to MAS, in terms of
serving as the foundation for inter-agent
communication on the one hand (Souza et al., 2016),
and knowledge bases that agents can use for
reasoning on the other hand (Laclavik et al., 2006;
Subercaze and Maret, 2011).
A plethora of MAS-specific methodologies and
tools has been developed over the past decades
(Sturm and Shehory, 2014). The use of ontologies in
MAS has been partially adressed in previous work on
MAS methodologies (Freitas et al., 2015). However,
the methodological development and utilization of
ontologies for reasoning and communication in MAS
remains an open issue as the ontological needs in
industrial applications cannot be met with existing
ontologies or even existing standardards,
notwithstanding their lack of formalization (Hodges
et al., 2017). Therefore, the authors developed the
Ontologies for Reasoning and Communication in
Multi-Agent Systems methodology (OReCo). It is not
designed to act as a standalone MAS development
1
http://jade.tilab.com
approach. Instead it can be utilized within existing
methodologies that follow a model-driven
engineering approach, e.g. (Chella et al., 2004;
Garcia-Ojeda et al., 2008; Linnenberg and Fay,
2018). Key features of OReCo are the integration
model-driven engineering artefacts, the utilization of
existing ontologies to reduce engineering effort, and
the integration into the popular MAS development
environment JADE
1
. OReCo comprises two parts:
ontology development and ontology utilization. The
former is the subject of this paper. The latter deals
with actual implementation tasks related to the
utilization of developed ontologies in MASs and will
be the subject of a future publication by the authors.
Henceforth, for the sake of brevity, we refer to the
term ontology as a machine-readable
conceptualisation of domain knowledge implemented
in the OWL 2 Web Ontology Language
2
(OWL 2), or
more specifically the OWL 2 DL profile. Note, that
this deviates from prominent definitions of the term,
which do not imply a representation language or
degree of formality, e.g. (Uschold and Gruninger,
1996). Only in some cases we use the term “ontology
in the broadest sense” to refer to the full spectrum of
conceptualisations as defined in (Lassila and
McGuinness, 2001). Occasionally, we use the terms
“lightweight ontoloy” and “heavyweight ontology” to
differentiate between human-centred UML class
2
http://www.w3.org/TR/owl2-overview/
450
Törsleff, S., Hildebrandt, C. and Fay, A.
Development of Ontologies for Reasoning and Communication in Multi-Agent Systems.
DOI: 10.5220/0008355804500457
In Proceedings of the 11th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management (IC3K 2019), pages 450-457
ISBN: 978-989-758-382-7
Copyright
c
2019 by SCITEPRESS Science and Technology Publications, Lda. All rights reserved
diagram representation and machine-readable
representation based on OWL 2 DL, respectively. The
communication and reasoning in MASs for cyber-
physical systems that are developed using model-
driven engineering methodologies will be referred to
as the “application domain” of the methodology from
here on. Specific desired solutions, e.g. a distribution
grid automation solution, we refer to as use cases.
Furthermore, a MAS developed for a specific use case
will be referred to as a system under development.
In an analogous manner, we use the term ontology
under development (also if multiple modular
ontologies are being developed at once).
The remainder of this paper is structured as
follows. In section 2 we present requirements
regarding the development of ontologies in the
application domain that have been gathered from
industry. Afterwards, in section 3, we provide an
analysis of related research and discuss in how far it
withstands the requirements we identified. The
methodology itself is presented in section 4 and
preliminary evaluation results regarding its
application in a smart energy grid-related research
project in section 5. Conclusions and an outlook on
future work of ours are provided in section 6.
2 REQUIREMENTS
The application domain outlined in the previous
section imposes a set of specific requirements which
a methodology, aiming to yield suitable ontologies,
has to satisfy. These requirements have been gathered
with involvement from industry partners in the
research project CrESt
3
(Collaborative Embedded
Systems), some of which originally were presented in
(Hildebrandt et al., 2018b), others were added in the
meantime.
R1 Systematic Requirements Elicitation:
Identifying requirements is the first step in ontology
development. As in software engineering in general,
requirements are a key determinant for successful
ontology development. Besides serving as the
foundation for designing and implementing an
ontology, structured requirements also facilitate
future reuse of the developed ontology. Therefore, the
methodology has to entail a systematic elicitation
process.
R2 Model-driven Engineering Artefact
Utilization: The methodology’s application domain
is the model-driven engineering of MASs. Thus,
diagrams that specify the message exchange and
3
https://crest.in.tum.de/
internal program logic of agents are available. These
are valuable artefacts with regard to the ontology
requirements and implicitly contain all intended uses
of the ontology under development. Accordingly, we
define the utilization of such artefacts as a
requirement.
R3 Ontology Modularity: Besides reusing existing
ontologies for a specific use case, it is desirable to
create ontologies that can be reused themselves.
Apart from systematic requirements documentation,
this goal can be achieved by modularizing the
ontology under development. The modularity also
facilitates the usage of distinct TBoxes for different
agents, thus keeping individual TBoxes lean.
Accordingly, ontology modularity represents another
requirement.
R4 Domain Expert Centricity: This requirement is
rooted in the fact that ontology experts usually do not
have the necessary knowledge to model the required
domain knowledge for a specific use case. Therefore,
domain experts who possess the necessary domain
knowledge have to take an active and central role in
ontology development.
R5 Intermediate Ontology Conceptualization:
While lightweight ontologies represented as UML
class diagrams are not as expressive as heavyweight
ontologies, they help to conceptualize domain
knowledge transparently. This facilitates the
integration of domain experts in ontology
development (see R4). For this reason, we define the
intermediate ontology conceptualization (with UML
class diagrams) as a requirement.
R6 Multi-stage Utilization of Existing Ontologies:
Domain conceptualization and ontology development
are time consuming tasks. In order to reduce the
engineering effort, it is hence advisable to make use
of existing ontologies (in the broadest sense)
whenever reasonable. This ranges from industry
standards as the basis for concept definitions to
integrating existing heavyweight ontologies. The
utilization of ontologies with varying degrees of
expressiveness requires their integration at different
stages of the development cycle.
R7 Defined Ontology Bridging Procedure: Aiming
for reusable, modular ontologies necessitates a
defined ontology bridging procedure to connect
ontologies as required by the system under
development.
Development of Ontologies for Reasoning and Communication in Multi-Agent Systems
451
3 STATE-OF-THE-ART
ANALYSIS
Existing ontology development methodologies can be
divided into three generations with increasing
sophistication and maturity (Simperl and Luczak-
Rösch, 2014). We limit our subsequent analysis to the
third generation, i.e. methodologies that emphasize
the collaborative nature of ontology development and
the involvement of non-ontology experts. Table 1
summarises the analysis of these methodologies with
regard to the requirements outlined in the previous
section.
DILIGENT is ontology expert-centric and focuses
community adaption of an initially-developed
ontology (Pinto et al., 2004). Another ontology
expert-centric methodology is HCOME (Vouros and
Kotis, 2006), which addresses the development of
“living ontologies” in the domain of knowledge-
intensive communities. In contrast, UPON Lite
integrates domain experts into the development (de
Nicola and Missikoff, 2016). Its drawback, however,
is its predominant use of tabular representations to
capture domain expert knowledge. R6 (Multi-stage
Utilization of Existing Ontologies) is only partially
satisfied by UPON Lite in that it recommends the
utilization of existing standards but does not address
their integration at multiple stages of the ontology
development. The innovation of the NeON
methodology lies in its emphasis of a “divide and
conquer” approach, in which the definition of
requirement subsets paves the way for developing and
utilizing ontology modules (de Figueroa Baonza,
2010). However, substantial domain expert
involvement is also missing in NeON.
OntologyMaturing in turn stresses domain expert
centricity and suggests capturing their knowledge
using lightweight ontologies, however it does not
satisfy, inter alia, the modularity requirement (Braun
et al., 2007). An approach that is based on established
methodologies from the software engineering domain
is given by SCIM (John et al., 2017), though it falls
short of addressing the integration of existing
ontologies at multiple stages of the development
cycle. The only approach that explicitly proposes the
utilization of UML diagrams for eliciting ontology
requirements has been presented in (Olszewska,
2015), though it is technically not a methodology and
also satisfies none of the other requirements.
Therefore, it is not included in Table 1.
As can be seen, there is no methodology available
that covers all or even the majority of the identified
requirements. This shortcoming was the basis for the
decision to develop a new ontology development
methodology.
Table 1: Requirements satisfaction of existing ontology
development methodologies.
R1
R2
R3
R4
R5
R6
R7
DILIGENT
X
X
HCOME
X
X
X
UPON Lite
X
X
(X)
NeON
X
X
X
X
Ontology
Maturing
X
X
SCIM
X
X
X
4 METHODOLOGY
As outlined in the introduction, the overall OReCo
methodology comprises two parts: ontology
development and ontology utilization. In this
contribution we focus on the ontology development
part, which is divided into three stages, each of which
comprising of multiple steps as depicted in Figure 1.
Indicated by the coloured dots, multiple roles are
involved in each step: the domain expert has deep
knowledge of the specific use case, was involved in
the design of the system under consideration and is
aware of relevant (de-facto) industry standards; the
software engineer is the end user of the ontology
insofar as utilizing it in the implementation of the
agents; the ontology expert is proficient in ontological
engineering and provides technical support.
The methodology presented hereafter builds on
previous publications of the authors (Hildebrandt et
al., 2018a, 2018b). While these publications were
rather generic in scope (covering, e.g., the
development of UML profiles), OReCo is limited to
the application domain as defined in section 1. This
allows omitting irrelevant steps of the generic
methodology and in turn emphasising aspects specific
to OReCo’s application domain.
OReCo employs a waterfall-like life-cycle model.
This is based on the rationale that the ontology’s
scope is determined by the design of the system under
development. That is, if the system design poses
specific requirements, the ontology has to satisfy
these requirements. An iterative-incremental
approach for the ontology development would thus be
inappropriate. This does of course not inhibit an
iterative-incremental approach being applied to the
overall system under development. In such cases,
multiple iterations of the OReCo methodology would
be driven by the enclosing development methodology
while itself remaining waterfall-like.
KEOD 2019 - 11th International Conference on Knowledge Engineering and Ontology Development
452
Figure 1: Overview of the OReCo methodology’s ontology development part.
Note, that OReCo does not prescribe a specific
OWL 2 profile. However, due to its wide tool support
and decidability (Hitzler et al., 2010), the OWL 2 DL
profile can be considered the default implementation
language for ontologies developed using OReCo.
4.1 Requirements Elicitation
The specification of ontology requirements is partly
based on the NeON methodology (Suárez-Figueroa et
al., 2009). Some of the tasks in the ontology
requirements specification according to NeON can be
omitted, since their results are the same for any use
case in OReCo’s application domain.
Table 2: Generic results of task 1-2 of the NeON Ontology
Requirements Specification for OReCo’s application
domain.
Purpose
Provide the foundation for
reasoning and communication in the
multi-agent system under
development
Scope
All classes and properties required
for communication and reasoning in
the system under development
Implementation
Language
Intermediate: UML class diagrams
Final: OWL
Intended End
Users
Software engineers (and, strictly
speaking, agents)
This holds for task 1 and 2, whose results are shown
in Table 2. Further tasks of the ontology requirements
specification according to NeON have been modified
to take into account specific needs of the application
domain and to improve overall validity with regard to
the requirements discussed in section 2.
Step 1.1. Annotate Competency Questions
processes artefacts from model-driven engineering,
i.e. the system design comprising UML diagrams
such as sequence diagrams (covering the message
exchange between agents) and activity diagrams
(covering the internal application logic of agents).
These diagrams are annotated with competency
questions (CQs) and answers using a custom UML
profile, which facilitates their automated export into
a tabular representation including a reference to the
respective agent. Strictly speaking, the annotation is
limited to informal CQs (Gruninger and Fox, 1995).
Later on, these informal CQs will be the basis for
formal SPARQL queries that are used for verification
once the ontology has been implemented. Note that
task 3 “Identify intended end uses” of the NeON
methodology has been omitted. This is due to the fact
that the intended uses are already given by the system
design.
Step 1.2 Identify Non-Functional Requirements
deals with aspects that are not covered by CQs, e.g. a
specific ontology naming convention or the
mandatory use of industrial standards. The latter
could result from the system design, e.g. if a agent
interacts with a technical system that requires the use
of a specific protocol. Non-functional requirements
are recorded in tabular format.
To lay the foundation for the reuse of existing
ontologies and the development of modular
ontologies, the requirements are clustered in step 1.3
Development of Ontologies for Reasoning and Communication in Multi-Agent Systems
453
Group Requirements. Useful groups are subdomains
such as technical properties of a battery storage or
generic concepts such as units of measure. Further
grouping may be performed in case of agents
requiring very specialized knowledge. Such groups,
by facilitating ontology segmentation, can help
reduce the TBox size (and thus inprove the reasoning
performance) of agents with distinct knowledge
bases. This step can also comprise the definition of a
group hierarchy, which determines the development
sequence in case the ontology under development is
composed of multiple interdependent ontologies.
Step 1.4 Validate Requirements serves to ensure
that the requirements are complete, consistent,
verifiable, understandable, unambiguous, concise and
modifiable. Guidance on how to evaluate these
criteria are provided in (Suárez-Figueroa et al., 2009).
An additional criterion is wether a requirement is
implementable, e.g.. if a CQ can be answered with an
ontology. This characteristic has been added since
non-ontology experts might add CQs that cannot be
covered by an ontology with reasonable effort. In case
of invalid requirements, step 1.1 and/or 1.2 of the
methodology have to be revisited (not depicted in
Figure 1). Note that the prioritization of requirements
as per the NeON methodology has been omitted. This
is based on the rationale that the ontology under
development serves as an enabler for reasoning and
communication in the system under development as
prescribed by the system design. This means that any
requirement that is not met by the ontology will have
negative repercussions with regard to the system
under development. Accordingly, it is assumed that
all requirements have equal priority. Potential
dependencies between ontology segments that
require a specific development sequence are
determined in step 1.3.
In step 1.5 Extract Glossary and Synonyms, the
foundation is laid for the subsequent analysis of
existing ontologies. The glossary comprises three
parts. The first part shows terms and frequency based
on the CQs, the second one based on the related
answers, and the third part identifies objects that can
be seen as instances of other classes. Additionally,
synonyms are recorded for each glossary term to
facilitate the survey of existing ontologies. The final
result of the requirements elicitation stage is an
Ontology Requirements Specification covering all
requirements for the ontology under development.
4.2 Conceptualization
The conceptualization stage begins with step 2.1
Analyse Existing Ontologies. In this step, the
ontology and domain experts first perform a survey to
identify suitable ontologies (in the broadest sense)
based on the glossary developed in step 1.5. This will
result in a pool of ontologies (in the broadest sense)
potentially ranging from plain dictionaries to
heavyweight ontologies. Ideally, the survey yields a
set of heavyweight ontologies that satisfy all
requirement (CQs and non-functional). In that case
one could jump straight to step 3.2 of the
methodology. The more realistic case, which we
assume henceforth, is the identification of both
lightweight and heavyweight ontologies, whereas
some CQs remain unanswered. These remaining CQs
will be processed in step 2.2 Implement Lightweight
Ontologies. Here, domain experts, supported by
ontology experts, conceptualize an ontology
addressing the remaining CQs as a UML class
diagram. We recommend performing this step with
the concept of so-called content ontology design
patterns in mind (Hitzler et al., 2016). These are small
modular and extendible ontologies which can be used
in multiple use cases. This approach is specifically
useful for industry standards that have a high
likelihood of being reused, but also in cases where
generic domain conceptualizations are useful. For
instance, consider a network topology ontology for
energy grids that can be extended for the domain gas,
electricity etc. Further orientation marks with regard
to the segmentation of the lightweight ontology
implementation task are given by the requirement
groups defined in step 1.3. Naturally, the pre-selected
lightweight ontologies need to be considered in this
step to avoid overlaps.
4.3 Implementation and Verification
The last stage of the development begins with step 3.1
Implement Heavyweight Ontologies. Inputs for this
step are the pre-selected and newly implemented
lightweight ontologies as well as the ontology
requirements specification. The latter provides
information not included in the lightweight
ontologies, e.g. naming conventions. The
groundwork for the implementation of content
ontology design patterns has been laid in step 2.2.
Thus, this step can be completed by sequential
transformation of each lightweight ontology into a
heavyweight ontology. Guidelines for the actual
modelling in OWL can be found in (Pinto et al.,
2004).
Figure 2 shows an exemplary scenario in which the
overall ontology under development comprises three
TBoxes that are being used by two distinct agents. In
this scenario, Agent1’s knowledge base imports
TBox1 and TBox2. Accordingly, “TBox bridges”, i.e.
KEOD 2019 - 11th International Conference on Knowledge Engineering and Ontology Development
454
TBox statements that connect existing TBoxes, are
required between TBox1 and TBox2 as well as
between TBox2 and TBox3. These statements are
symbolized by the white lines within the agent’s
knowledge bases. Technical alternatives and
guidance on how to bridge ontologies can be found in
(Hodges et al., 2017).
Figure 2: Exemplary utilization and bridging of TBoxes in
different agents.
Step 3.2 Bridge Heavyweight Ontologies deals with
integrating the resulting heavyweight ontologies. A
key determinant of this step is the utilization of these
ontologies by the agents, which is captured in the
ontology requirements specification.
Lastly, in step 3.3 Verify Heavyweight Ontologies
the developed heavyweight ontologies are verified.
First, a consistency check of the resulting ontology is
performed using a suitable reasoner, e.g. HermiT
4
. If
this is successful, SPARQL queries are defined for
the CQs identified earlier. The results of the query
need to be consistent with the respective answers of
each CQ in order for this step to succeed. Otherwise,
step 3.1 or 3.2 have to be revisited.
5 EVALUATION
This section deals with a preliminary evaluation of
the proposed methodology and its satisfaction of the
requirements presented in section 2. The use case for
the evaluation was a MAS-based low-voltage
distribution grid automation solution that had been
developed in Agent.HyGrid
5
, a research project
dealing with smart energy grids. UML diagrams
specifying the message exchange between agents and
their inter program logic, i.e. the system design,
formed the basis for the ontology development,
illustrating the satisfaction of requirement R2. It was
possible to group the identified CQs based on their
attribution to agents. For instance, one group
comprised CQs focussing the grid topology and was
only relevant to one agent, while another group
adressed the flexibility of distributed energy
resources, which multiple agents exchange
4
http://www.hermit-reasoner.com
information about. This provided the foundation for
the concurrent execution of subsequent
conceptualisations and underlines the methodology’s
satisfaction of R1. As no suitable ontology for
flexibilities could be identified by the domain experts
in step 2.1, they modelled a lightweight ontology
from scratch, based on the unanswered CQs (R4 and
R5). In contrast, the IEC 61970-501 standard was
identified as suitable for conceptualizing the grid
topology, which conveniently had been implemented
as a heavyweight ontology already (Schumilin et al.,
2017). After the lightweight ontology for flexibilities
had been implemented as a heavyweight ontology in
step 3.1, it was bridged with the grid topology
ontology in step 3.2 (R6 and R7). While the
flexibility ontology is now used by all agents, the grid
topology is only used by one agent, which
demonstrates the advantageuousness of the ontology
modularity support by our methodology (R3).
6 CONCLUSIONS AND
OUTLOOK
Ontologies have the potential to significantly reduce
the engineering effort required in MAS development.
OReCo is a methodology comprising two parts:
developing ontologies and utilizing ontologies for
reasoning and communication in MASs for cyber-
physical systems. In this contribution, we presented
the first part, which is unique in addressing
requirements specific to this application domain.
While providing significant value in its current state
already, additional research is required to improve
usability and reduce the required engineering effort
even further.
In this spirit, the authors are currently conducting
a systematic survey of existing ontologies in the smart
grid space. The results of this survey will serve as the
foundation for a software tool that supports ontology
and domain experts in the development process by
providing recommendations for reuse of existing
ontologies based on the glossary in step 2.1 Analyse
Existing Ontologies.
Further improvements to the methodology will
also result from ongoing and planned evaluation
activity in the context research projects dealing with
smart factories and energy grids.
5
http://www.agent-hygrid.net
Development of Ontologies for Reasoning and Communication in Multi-Agent Systems
455
ACKNOWLEDGEMENTS
The contribution presented in this paper was funded
by the German Federal Ministry of Education and
Research under grant number 01IS16043Q
Collaborative Embedded Systems (CrESt).
REFERENCES
Braun, S., Schmidt, A., Walter, A., 2007. Ontology
Maturing: a Collaborative Web 2.0 Approach to
Ontology Engineering, in: Proceedings of the
Workshop on Social and Collaborative Construction of
Structured Knowledge (CKC 2007) at the 16th
International World Wide Web Conference
(WWW2007).
Chella, A., Cossentino, M., Sabatucci, L., 2004. Tools and
Patterns in Designing Multi-Agent Systems with
PASSI. WSEAS Trans. Commun. 3, 352358.
de Figueroa Baonza, M. del C.S., 2010. NeOn Methodology
for Building Ontology Networks: Specification,
Scheduling and Reuse.
de Nicola, A., Missikoff, M., 2016. A lightweight
methodology for rapid ontology engineering. Commun.
ACM 59, 7986. https://doi.org/10.1145/2818359
Freitas, A., Bordini, R.H., Meneguzzi, F., Vieira, R., 2015.
Towards integrating ontologies in multi-agent
programming platforms, in: Proceedings of the
IEEE/WIC/ACM International Joint Conference on
Web Intelligence and Intelligent Agent Technology
(WI-IAT 2015). IEEE, pp. 225226. https://doi.org/
10.1109/WI-IAT.2015.207
Garcia-Ojeda, J.C., DeLoach, S.A., Robby, Oyenan, W.H.,
Valenzuela, J., 2008. O-MaSE: A Customizable
Approach to Developing Multiagent Development
Processes, in: Luck, M., Padgham, L. (Eds.), Agent-
Oriented Software Engineering VIII: 8th International
Workshop, AOSE 2007. Springer Berlin Heidelberg,
Berlin, Heidelberg. https://doi.org/10.1007/978-3-540-
79488-2_1
Gruninger, M., Fox, M.S., 1995. Methodology for the
Design and Evaluation of Ontologies, in: Workshop on
Basic Ontological Issues in Knowledge Sharing
(IJCAI’95).
Hildebrandt, C., Törsleff, S., Bandyszak, T., Caesar, B.,
Ludewig, A., Fay, A., 2018a. Ontology engineering for
collaborative embedded systems - Requirements and
initial approach. CEUR Workshop Proc. 2060, 5766.
Hildebrandt, C., Törsleff, S., Caesar, B., Fay, A., 2018b.
Ontology Building for Cyber-Physical Systems : A
domain expert-centric approach, in: 14th IEEE
International Conference on Automation Science and
Engineering (IEEE CASE 2018).
Hitzler, P., Gangemi, A., Janowicz, K., Krisnadhi, A.A.,
Presutti, V., 2016. Ontology Engineering with
Ontology Design Patterns: Foundations and
Applications. IOS Press.
Hitzler, P., Krötzsch, M., Rudolph, S., 2010. Foundations
of Semantic Web Technologies. Chapman & Hall.
Hodges, J., García, K., Ray, S., 2017. Semantic
Development and Integration of Standards for
Adoption and Interoperability. Computer (Long.
Beach. Calif). 50, 2636.
John, S., Shah, N., Stewart, C., Samlov, L., 2017. Software
Centric Innovative Methodology for Ontology
Development, in: Proceedings of the 9th International
Conference on Knowledge Engineering and Ontology
Development (KEOD 2017). pp. 139146.
https://doi.org/10.5220/0006482901390146
Laclavik, M., Balogh, Z., Babik, M., Hluchy, L., 2006.
AgentOWL: Semantic Knowledge Model and Agent
Architecture. Comput. Informatics 25, 419437.
Lassila, O., McGuinness, D.L., 2001. The Role of Frame-
Based Representation on the Semantic Web, Technical
Report KSL-01-02. https://doi.org/10.1177/0363
54659702500204
Leitao, P., Karnouskos, S., Ribeiro, L., Lee, J., Strasser, T.,
Colombo, A.W., 2016. Smart Agents in Industrial
Cyber Physical Systems. Proc. IEEE 104, 10681101.
https://doi.org/10.1109/JPROC.2016.2521931
Linnenberg, T., Fay, A., 2018. Software Engineering for
Agent Based Energy Systems, in: 14th IEEE
International Conference on Automation Science.
Olszewska, I.J., 2015. UML Activity Diagrams for OWL
Ontology Building, in: Proceedings of the 7th
International Joint Conference on Knowledge
Discovery, Knowledge Engineering and Knowledge
Management (IC3K2015) - Volume 2: KEOD. pp. 370
374. https://doi.org/10.5220/0005633103700374
Pinto, H.S., Staab, S., Tempich, C., 2004. DILIGENT:
Towards a fine-grained methodology for DIstributed,
Loosely-controlled and evolvInG Engineering of
oNTologies, in: Proceedings of the 16th European
Conference on Artificial Intelligence (ECAI’04).
Schumilin, A., Stucky, K.U., Sinn, F., Hagenmeyer, V.,
2017. Towards ontology-based network model
management and data integration for smart grids, in:
2017 Workshop on Modeling and Simulation of Cyber-
Physical Energy Systems (MSCPES).
https://doi.org/10.1109/MSCPES.2017.8064530
Simperl, E., Luczak-Rösch, M., 2014. Collaborative
ontology engineering: A survey. Knowl. Eng. Rev. 29,
101131. https://doi.org/10.1017/S0269888913000192
Souza, M., Moreira, A., Vieira, R., Meyer, J.-J.C., 2016.
Integrating Ontology Negotiation and Agent
Communication, in: Tamma, V., Dragoni, M.,
Gonçalves, R., Ławrynowicz, A. (Eds.), Ontology
Engineering. Springer International Publishing, Cham,
pp. 5668.
Sturm, A., Shehory, O., 2014. The Landscape of Agent-
Oriented Methodologies, in: Shehory, O., Sturm, A.
(Eds.), Agent-Oriented Software Engineering -
Reflections on Architectures, Methodologies,
Languages, and Frameworks. Springer-Verlag Berlin
Heidelberg, pp. 137154. https://doi.org/10.1007/978-
3-642-54432-3
KEOD 2019 - 11th International Conference on Knowledge Engineering and Ontology Development
456
Suárez-Figueroa, M.C., Gómez-pérez, A., Villazón-
Terrazas, B., 2009. How to Write and Use the Ontology
Requirements Specification Document How to Write
and Use the Ontology Requirements, in: Meersman, R.,
Dillon, T., Herrero, P. (Eds.), On the Move to
Meaningful Internet Systems. Springer-Verlag, Berlin
Heidelberg, pp. 966982. https://doi.org/10.1007/978-
3-642-05151-7
Subercaze, J., Maret, P., 2011. Programming Semantic
Agent for Distributed Knowledge Management, in:
Elci, A., Koné, M.T., Orgun, M.A. (Eds.), Semantic
Agent Systems - Foundations and Applications.
Springer-Verlag Berlin Heidelberg, pp. 4765.
Uschold, M., Gruninger, M., 1996. Ontologies : Principles
, Methods and Applications. Knowl. Eng. Rev. 11, 93
136. https://doi.org/10.1.1.111.5903
Vouros, G.A., Kotis, K., 2006. Human-centered ontology
engineering: The HCOME methodology. Knowl. Inf.
Syst. 10, 109131. https://doi.org/10.1007/s10115-005-
0227-4
Vrba, P., Marik, V., Siano, P., Leitao, P., Zhabelova, G.,
Vyatkin, V., Strasser, T., 2014. A review of agent and
service-oriented concepts applied to intelligent energy
systems. IEEE Trans. Ind. Informatics 10, 18901903.
https://doi.org/10.1109/TII.2014.2326411
Development of Ontologies for Reasoning and Communication in Multi-Agent Systems
457