Agent-based Electronic Commerce with Ontology Services and Social
Network based Support
Virgínia Nascimento, Maria João Viamonte, Alda Canito and Nuno Silva
GECAD, Knowledge Engineering and Decision Support Research Center, Institute of Engineering,
Polytechnic of Porto (ISEP/IPP), Porto, Portugal
Keywords: Agent Mediated Electronic Commerce, Semantic Heterogeneity, Ontology Matching, Ontology Alignment
Negotiation, Emergent Social Networks.
Abstract: In this paper we approach the semantic heterogeneity problem which arises in agent mediated electronic
commerce, presenting the AEMOS system as a promising answer. AEMOS is an agent mediated electronic
commerce platform which main goal is to enable an efficient and transparent negotiation between agents
even when they use different ontologies to represent the same domain of knowledge. The system provides
ontology services, more specifically ontology matching services, which are improved by the exploitation of
emergent social networks. In this paper we present the currently implemented system and demonstrate how
an agent mediated electronic commerce system may benefit from the inclusion and combination of ontology
services and social network based support.
Electronic commerce (e-commerce) is a widely used
technology with an increasing popularity in today’s
business (Du et al., 2005). In this type of commerce
the information becomes more easily available,
increasing the possibility of achieving more
satisfactory deals. However, the amount of available
information also becomes a problem, being difficult
for a human user to compare all possible deals in
order to achieve the best one.
Intelligent agents present characteristics that
make them a powerful tool to overcome this
problem. However, the diversity of the available
information, which is normally represented for
human comprehension only, turns the development
of fully automated systems into a challenge.
In order to overcome this problem, ontology
centered approaches have been proposed (Mei et al.,
2009, Cao et al., 2009) and e-commerce key players
such as Google, Yahoo, Amazon and O’Reilly are
progressively supporting these through micro-
formats, micro-data and RDFa (O'Brien, 2009).
However, given the natural diversity of such an
open and accessible environment, the involved
entities may possess different conceptualization
about their needs and capabilities, giving rise to a
semantic heterogeneity problem that is seen as a
corner stone for agents’ interoperability.
Based on these issues we developed AEMOS –
Agent-based Electronic Market with Ontology
Services (Nascimento et al., 2012, Viamonte et al.,
2012, Viamonte et al., 2011, Silva et al., 2009), an
innovative project (PTDC/EIA-EIA/104752/2008)
supported by the Portuguese Agency for Scientific
Research (FCT).
The main goal of this project is to provide an
agent mediated e-commerce (AMEC) platform
capable of enabling an efficient and transparent
negotiation between agents even when they use
different ontologies, ensuring that they are able to
understand each other and correctly assess the terms
and conditions of each transaction. For that we
follow an ontology based information integration
approach, exploiting the ontology matching
paradigm (Euzenat and Shvaiko, 2007), which is
improved by the application and exploitation of
emergent social networks (SN).
During the development of this project different
models have been proposed and tested, e.g. see
(Nascimento et al., 2012, Viamonte et al., 2012,
Viamonte et al., 2011, Silva et al., 2009). In this
paper we present the currently implemented system,
clarifying the fundaments leading to our choices,
and presenting the more recently achieved results.
Nascimento V., João Viamonte M., Canito A. and Silva N..
Agent-based Electronic Commerce with Ontology Services and Social Network based Support.
DOI: 10.5220/0004448904970504
In Proceedings of the 15th International Conference on Enterprise Information Systems (ICEIS-2013), pages 497-504
ISBN: 978-989-8565-60-0
2013 SCITEPRESS (Science and Technology Publications, Lda.)
We start by briefly describing the research
background (Section 2). Then we present an
overview of the AEMOS system (Section 3),
highlighting the implemented ontology services
(Section 4) and SN-based support (Section 5). In
order to assess our proposal we present details about
the AEMOS prototype and describe some
experiments analyzing the achieved results (Section
6). Finally we draw some conclusions and suggest
follow-up research efforts (Section 7).
The most frequent approaches for AMEC systems
consider simplified and limited solution in order to
avoid semantic problems. Some consider the
existence of a commonly agreed ontology, such that
to participate in the market each agent has to adopt
this ontology, e.g. (Viamonte et al., 2007). Other
approaches, consider the existence of different
ontologies, but only allow communication between
agents that use the same ontology, e.g. (Cui-Mei,
The Foundation for Intelligent Physical Agents
(FIPA) has analyzed the interoperability problem in
heterogeneous multi-agent systems (MAS) and has
proposed the introduction of an agent that, among
other responsibilities, would be capable of
translating expressions between different ontologies
(FIPA, 2001). An implementation of such an agent
is proposed in (Briola et al., 2008), where the
translation service is achieved through ontology
Ontology matching (Euzenat and Shvaiko, 2007)
can be described as the process of discovering
semantic relations (i.e. correspondences) between
the concepts and properties of two ontologies. The
discovered relations are represented in an ontology
alignment document so they can be applied in data
There are already some approaches for AMEC
systems which include ontology matching services
in order to solve semantic problems. Some examples
are the models presented in (Malucelli et al., 2006)
and in (Wang et al., 2010). However, each approach
tends to focus on a particular aspect or phase of the
known behavior models, and often ignore the effect
that the complexity of the matching processes can
have in the communications’ efficiency.
Ontology matching is a naturally ambiguous and
subjective process, leading to different alignments
that may be more or less adequate to each
negotiation and therefore influence its efficiency and
result. The quality and adequacy of an ontology
alignment is very important in the negotiation, since
it may determine the efficiency of the interaction.
For example, a consumer may request a product that
a supplier has on its inventory, but by using an
inadequate alignment, relevant information may be
lost during the transformation process causing the
supplier not to be able to match it.
On the other hand, detecting incorrect or
inadequate ontology alignments is not a trivial task.
A negotiation may fail because the alignment is
inadequate to the current context, but it can also fail
simply because the supplier does not have the
desired products, or even because the agents have
different goals (e.g. conflicting prices).
The AEMOS system is based on ISEM (Viamonte et
al., 2007), which is an agent-based simulator system
for e-commerce, originally developed for studying
agents’ market strategies. In reality, AEMOS is an
evolution of the ISEM system, keeping all its
original functionalities, but allowing agents to use
different ontologies.
AEMOS provides ontology services in order to
enable negotiation between agents that use different
ontologies. The system exploits the ontology
matching paradigm (Euzenat and Shvaiko, 2007),
selecting and suggesting possible alignments
between the agents’ ontologies, and letting the
agents choose which one should be used to translate
the subsequent exchanged messages.
In order to overcome issues related to how the
chosen ontology alignment may influence the
business negotiation efficiency, the relevance of
social network analysis (SNA) (Wasserman and
Faust, 1994) in recommending ontology alignments
for e-commerce negotiations is claimed, by
including in the system a SN-based support
component, capable of improving the ontology
alignments recommendations and supporting the
agents’ decisions about which alignment to choose.
3.1 Multi-Agent Model
The AEMOS multi-agent model, illustrated in
Figure 1, includes several types of agents classified
in two main categories, namely: the business agents
and the supporting agents.
Figure 1: AEMOS multi-agent model.
The business agents represent real world entities,
which possess business goals to be satisfied. These
agents are highly customizable and dynamic. In each
situation, they adapt their strategies, according to the
present context and based on their updated
knowledge (Viamonte et al., 2007). Currently two
types of business agents are considered, namely:
Buyer (B) – agent representing a consumer,
i.e., an entity, normally a person wishing to
acquire a set of products;
Seller (S) – agent representing a supplier, i.e.
an entity, usually a company wishing to sell a
set of products.
The supporting agents are those providing
services that allow business agents to carry
transactions with each other in order to satisfy their
goals. This category can be further divided in two
groups, namely: the service intermediary agents and
the system management agents.
The service intermediary agents support the
business agents, providing services that enable an
efficient interoperability between them. In this
category are the agents:
Market Facilitator (MF) – agent that
coordinates the interaction between business
agents, being responsible for ensuring that the
communicating agents are able to understand
each other. When a B agent is registered, a
MF agent is associated; from that moment on,
all messages related to the business
negotiation process pass through the
associated MF agent;
Ontology Matching intermediary (OM-i) –
agent responsible for the ontology services,
recommending possible ontology alignments
for each business negotiation, and
transforming the exchanged messages
according to the agreed alignment. When a
MF agent is initiated an OM-i agent is
associated; from that moment on, all the
requests related to ontology matching services
are sent to the associated OM-i agent;
Social Network intermediary (SN-i) – agent
responsible for the SN-based support,
providing advice about the adequacy of the
ontology alignments to each business
negotiation. When an OM-i agent is initiated,
or a business agent registers in the market, a
SN-i agent is associated; from that moment
on, all requests related to SN-based support
are sent to the associated SN-i agent.
Normally there are multiple agents of these types
per marketplace, being initialized when necessary.
The system management agents are responsible
for granting the system’s dynamism, flexibility and
correct functioning. In this category are the agents:
Market Manager (MM) – agent responsible to
manage all supporting agents and to register
business agents so they can participate in the
market. Normally there is only one agent of
this kind per marketplace;
Market Extension Manager (MEM) – agent
that aids the MM on its functions, allowing the
dynamic addition of machines where
supporting agents may be initialized. The
presence of this kind of agent is optional,
although normally there are multiple agents of
this type per marketplace;
Market Data Manager (MDM) – agent that
collects and maintains information about the
market participants and their activities, writing
statistical reports which allow evaluating the
system’s performance. Normally there is only
one agent of this type per marketplace;
Clock – agent that simulates the evolution of
time, notifying the appropriate agents about
periodic (or scheduled) events. Normally there
is only one agent of this type per marketplace.
In addition to these types of agents there are the
Communication Facilitator agents, which are
responsible for establishing communications
between the different agents of the system. Since our
system is based in OAA (OAA, 2001), this role is
played by the Facilitator agent provided by this
3.2 Interaction Protocol
To participate in the market, the business agents
must register first, providing information about the
ontologies that they use, and sharing (parts of) the
profile of the entity they represent. This information
is stored by MF and SN-i agents. Once registered,
the agents are allowed to negotiate. For that, B
agents start announcing their buying products and
wait for S agents to formulate proposals. Figure 2
illustrates the interactions between the main
intervenient during a business negotiation.
Figure 2: General agents’ interaction.
When the negotiation starts, the responsible MF
selects the S agents that might be able to satisfy the
B agent’s request. For that it follows an ontology-
based approach, selecting: (i) the S agents that use
the same ontology as the B; and, (ii) supported by an
OM-i, the ones that use ontologies that can be
aligned with it. Therefore, the business negotiations
may occur in two different scenarios:
A scenario where both agents use the same
ontology – the MF acts as a proxy between B
and S, simply receiving and forwarding
A scenario where the agents use different
ontologies – it is necessary to find an
agreement about the alignment between the
respective ontologies that should be used to
translate the exchanged messages. For that the
MF requests an OM-i to mediate an ontology
alignment negotiation between B and S. If an
agreement is achieved, the subsequent
exchanged messages are sent to the OM-i,
which translates their content according to the
agreed alignment ensuring that the message
receiver will be able to understand it.
During the business negotiation the involved
agents, B and S, exchange proposals and
counterproposals, following a protocol based on the
FIPA’s “Iterated Contract Net Interaction Protocol
Specification” (FIPA, 2002), terminating the
negotiation when an agreement is achieved or when
they have no more proposals to formulate.
When a business agent satisfies all its business
goals, or its deadlines are reached, it must terminate
its activity, notifying the market and declaring the
achieved results.
3.3 Ontology Alignment Negotiation
The ontology alignment negotiation initiates when a
MF sends a request to an OM-i identifying (i) both
agents, (ii) the respective ontologies and (iii)
providing information about the B agent’s request.
The OM-i selects, from its repository, all the
possible alignments between the indicated
ontologies. Then, it performs sorting and filtering
actions, following its internal criteria and/or
requesting a SN-i to rank the alignments, obtaining a
list of possible alignments and their respective score.
Both B and S, analyze the recommended alignments
taking into account their preferences, replying to the
OM-i with the list of the alignments which they
consider acceptable.
The OM-i analyzes both replies and checks if
there is an agreement, i.e., if some alignment was
selected by both agents. If there is no agreement,
depending on the system configuration, the
negotiation may terminate, or proceed, with the OM-
i refining its list of recommended alignments and
asking agents to reconsider their options and criteria.
Otherwise, if there is an agreement, the OM-i
notifies both agents and the MF about the agreement
and proceeds with the transformation of the B
agent’s request. From that moment on, all the
subsequent exchanged messages between the agents
are forward to the OM-i for transformation.
The ontology services are provided by OM-i agents.
An OM-i is responsible for (i) discovering
ontologies and ontology alignments, (ii) providing
information about ontologies and alignments, (iii)
proposing alignments for negotiation, (iv)
coordinating alignment negotiations, and (v)
transforming ontology’s instances when requested.
Although these responsibilities are attributed to
the OM-i, this agent normally requests services from
other specialized agents in order to perform these
tasks (especially for the ontology matching process),
e.g. see (Viamonte et al., 2011).
To improve performance, currently, the ontology
matching process is performed externally to the
negotiation process. It is considered a registry of
ontologies that are recognized by the agent and a
repository of possible alignments between them.
This information can be updated at any time, as new
ontologies are discovered and ontology alignments
are created.
We also consider that agents may represent their
domain of knowledge using public ontologies, i.e.
ontologies publicly accessible, having their own
URL, either through a dedicated web page or being
stored in web repositories. Therefore it is possible to
gather ontologies used in an e-commerce context
and discover possible alignments between them, or
even collect already existent alignments from public
web sources.
When the alignment negotiation is requested, the
OM-i selects from its alignments repository the ones
that involve both of the indicated ontologies. It then
ranks, sorts and filters the alignments either by (i)
requesting a SN-i to rank the alignments for the
business negotiation, or (ii) by analyzing their
coverage of the ontology’s concepts and properties
used by the B to describe the requested product. It
then coordinates the alignment negotiation following
the protocol previously described (cf. Section 3.3).
In order to improve its recommendations, in each
alignment negotiation’s iteration, the OM-i stores
and maintains information about the recommended
alignments and the achieved agreements.
The transformation of a message’s content (i.e.
ontology’s instance) is performed using the agreed
alignment. This process is provided by information
integration tools such as MAFRA Toolkit (Maedche
et al., 2002) and it is transparent to the agents.
The SN-based support is introduced in the system in
order to enhance the communication’s efficiency, by
improving the evaluation of the alignments’
adequacy to each business negotiation.
In previous work, we proposed two different
models for this component: one based in explicit
social networks (Viamonte et al., 2012) and another
based in emergent social networks (Nascimento et
al., 2012). In the first the business agents provide
information about their own evaluations of their
business partners and used ontology alignments,
while in the latter the SN-i agents analyze the
similarities between the agents’ profiles and
behavior, as well as the outcomes of their
interactions. We consider this last model more
interesting and adequate to our current problem
since it is less demanding for the business agents
(increasing the transparency of the process), and less
dependent on them as well, allowing us to overcome
problems such as the feedback credibility (Das et al.,
2011). This model also allows us to take advantage
of the collaboration between the system’s agents.
During the market activity, the SN-i collects
information about its participants and their
interactions. Then it builds and maintains the
relationship graph, applying SNA techniques
(Wasserman and Faust, 1994) in order to capture
proximity relations between agents, and adequacy
relations from alignments to agents, which emerge
during the agents’ activities in the market. By
combining this information, the SN-i is able to
evaluate the adequacy of the alignments to each
business negotiation.
A detailed description of the SN-i agent’s model,
as well as the fundaments behind it, can be found in
(Nascimento et al., 2012). In this paper we present
only the key aspects of its responsibilities, which are
as follows.
Collect information throughout the market: the
SN-i receives information from the other agents on
the market which will allow it to, in return, support
them in their tasks, e.g.: (i) business agents provide
information about the profile of the entities they
represent and about ontologies’ usage and
preferences; (ii) MF agents provide information
about the business negotiations between the agents
(e.g. both agents’ identification, the used alignments,
the negotiation outcome, the satisfaction of B with
the deal); and (iii) OM-i agents provide information
about ontologies, alignments and previous alignment
Evaluate Agent-To-Agent proximity: based in
some theories supported in the literature, the SN-i
combines a series of factors in order to capture
proximity relations between agents. These factors
are: (i) the similarity between the agents’ profiles
and ontologies usage and preferences; (ii) the
similarity between their interactions with other
agents; (iii) the success rate of their own previous
business negotiations; and (iv) the average
satisfaction of B about purchased products from S.
Evaluate Alignment-To-Agent adequacy: to
determine the adequacy of an alignment to an agent,
the SN-i evaluates: (i) the alignment’s coverage of
the agent’s used ontologies’ concepts and properties
(considering their respective relevance); (ii) the
agent’s success rate in business negotiations using
the alignment; and (iii) the agent’s average
satisfaction in deals using the alignment.
Evaluate Alignment-To-Business-Negotiation
adequacy: the adequacy of an alignment to a
business negotiation will depend on many factors,
namely: (i) the coverage of the alignment in relation
to the requested product’s description; (ii) the
general success rate in negotiations using the
alignment; (iii) the general average satisfaction in
deals using the alignment; (iv) the adequacy of the
alignment to each of the involved agents; and (v) the
adequacy of the alignment to the agents closest (i.e.
with high proximity relations) to the involved
In order to test and validate the AEMOS proposed
model a new system was developed and several
experiments were performed considering different e-
commerce scenarios. In this section we present
details about the implemented system and analyze
the achieved results.
6.1 The AEMOS Prototype
The AEMOS system was developed in Open Agent
Architecture (OAA). The OAA’s Interagent
Communication Language is the interface and
communication language shared by all agents, and
each agent is implemented in Java. The model can
be distributed over a network of computers, which is
a very important advantage considering the large
amount of agents that may exist per market.
In order to test and validate the AEMOS model,
and compare it with other AMEC approaches, we
developed an application that enables the simulation
of different e-commerce scenarios. The AEMOS
simulator is very flexible as it allows defining the
model to simulate, including the available services’
configuration, the number of business agents, each
agent’s type, ontologies and strategies. The set-up of
the AEMOS system is characterized by three
dimensions: (i) the business agents’ dimension,
which includes the business entities’ profiles,
inventories/shopping lists and satisfaction measuring
functions; (ii) the ontology services dimension,
which includes the considered ontologies and
alignments; and (iii) the SN dimension, which
includes the SN-i agents’ parameters to capture the
emergent SN and perform evaluations. Each of these
dimensions includes several parameters which can
be configured.
6.2 Case Study
In this experiment we intend to demonstrate how the
inclusion of ontology services, and combination
with SN-based support, can improve the business
negotiation’s efficiency, leading to a higher
satisfaction in the performed transactions.
Due to lack of space, the used configuration is
not fully presented here. We describe only key
aspects in order to provide a better understanding of
the achieved results.
We study a simple market composed by 4
suppliers and 7 consumers, whose profiles are
randomly generated. In order to correctly assess our
proposal the agents negotiate the same type of
product. To demonstrate the usefulness of our
system we include situations where agents that use
different ontologies could provide more satisfactory
deals, i.e. supply products with more similar
characteristics to the ones desired by the B, or that
contemplate the ones that the B values most.
Note that in this experiment we are focusing on
the ontology dimension of the negotiation, so other
factors in the formulation/selection of a proposal
(e.g. price, delivery time, quality of service) are
considered to be similar and compatible for each
agent. Therefore, the satisfaction of a B with a deal
will correspond to the similarity between the
purchased product and the desired one. The
satisfaction value is obtained by averaging each
attribute’s similarity, weighing by the relevance that
B attributes to each attribute. More details on this
function can be found in (Nascimento et al., 2012).
We consider three different ontologies. For each
pair of ontologies, two alignments were specified: (i)
one containing all the correct correspondences
between the ontologies; and (ii) another containing
less of these correct correspondences and including
some others which are incorrect.
6.3 Scenarios and Results
In order to test and validate our model, and based in
the most frequent approaches for AMEC (cf. Section
2), we considered four different scenarios:
Scenario 1 – A scenario without both ontology
services and SN-based support. The agents
negotiate only with agents that use the same
Scenario 2 – A scenario with ontology
services but no SN-based support. However,
this scenario only considers the correct
alignments. The OM-i agents evaluate the
alignments’ coverage of the concepts and
properties used by the B to describe the
requested product, and the agents choose the
ones with higher coverage;
Scenario 3 – A scenario similar to scenario 2
but considering all created alignments;
Scenario 4 – A scenario with both ontology
services and SN-based support. OM-i request
an SN-i to evaluate the alignments and the
agents choose the ones with higher score.
Each scenario ran several times in the AEMOS
simulator. Table 1 presents the average satisfaction
in deals obtained in each scenario, as well as the
average adequacy of the used alignment.
Table 1: Average results from each scenario.
in Deals
Scenario 1
0.55 -
Scenario 2
0.65 -
Scenario 3
0.52 0.14
Scenario 4
0.62 0.30
Comparing the first two scenarios allows us to
demonstrate how the system could benefit from the
inclusion of the ontology services by itself.
However, while in the first scenario the agents are
limited to communicate with agents that use the
same ontologies, the second represents an unrealistic
situation, by considering that the ontology
alignments are always semantically correct and
equally adequate. When we include the incorrect
alignments in the system the achieved satisfaction in
deals is even lower than the one achieved in the first
scenarios. This is due to the fact that, in this third
scenario the agents will continue choosing the less
adequate alignments which will cause a severe
impact on their business satisfaction.
Figure 3: Trending of satisfaction in deals and adequacy of
the used alignment in scenarios 3 and 4.
As Figure 3 illustrates, by including the SN-
based component, the alignment recommendations
tend to improve with time, allowing the agent to
choose the most adequate alignments achieving a
higher business satisfaction.
Moreover, by comparing the results achieved in
scenarios 3 and 4, we can conclude that, when the
SN-based support is included the agents need to
negotiate less (20%) to achieved their business
goals, there are less (29,4%) failed interactions, and
there are more (14.3%) transacted products.
The exploitation of the ontology matching paradigm
has been proposed in order to overcome the
semantic heterogeneity problem which arises in
open MAS. However, ontology matching may turn
into a highly complex and time consuming process
affecting the system’s performance. Ontology
matching is also a naturally ambiguous and
subjective process, which may lead to different
alignments that may be more or less adequate to
each negotiation, affecting its efficiency and result.
On the other hand, detecting incorrect or inadequate
alignments is not a trivial task due to the different
variables that may contribute for the negotiation
The system presented in this paper takes these
issues into account, providing an AMEC system
capable of enabling an efficient and transparent
negotiation between agents, even when they use
different ontologies. For that the system includes
and combines ontology matching services and SN-
based support. The ontology services allow agents to
interact with a higher range of business partners,
increasing the probability of achieving more
satisfactory deals, while the SN-based component
improves the business negotiations’ efficiency by
improving the recommendation/usage of ontology
The performed experiments have demonstrated
the usefulness and effectiveness of the implemented
model, being successful in the fulfilling of our initial
goals. However, we believe there are some aspects
in the systems which can be improved and future
research directions can be referred. For example, the
SN-based support component could be significantly
improved, exploiting other SNA techniques, in other
to achieve a more sophisticated model. Another
aspect to improve soon is the negotiation protocol,
both for business and ontology alignment
negotiations. The currently implemented protocol is
a legacy from the ISEM system and we believe that
a more sophisticated/efficient protocol could be
This work is supported by FEDER Funds through
the “Programa Operacional Factores de
Competitividade - COMPETE” program and by
National Funds through FCT “Fundação para a
Ciência e Tecnologia” under the projects: FCOMP-
Briola, D., Locoro, A. & Mascardi, V. 2008. Ontology
Agents in FIPA-compliant Platforms: a Survey and a
New Proposal. In: Baldoni, M., Cossentino, M., De
Paoli, F. & Seidita, V. (eds.) Atti del Workshop Dagli
Oggetti agli Agenti , WOA'08. Seneca Edizioni.
Cao, M., Feng, Y. & Liu, Z. 2009. E-Commerce Oriented
Negotiating Agent Communication Model. 42nd
Hawaii International Conference on System Sciences
(HICSS '09). Waikoloa, Big Island, Hawaii, USA:
IEEE Computer Society.
Cui-Mei, B. 2009. Combining Intelligent Agent with the
Semantic Web Services for Building An e-Commerce
System. 2009 IEEE International Conference on E-
Business Engineering (ICEBE '09). Macau, China:
IEEE computer society.
Das, A., Islam, M. M. & Sorwar, G. 2011. Dynamic Trust
Model for Reliable Transactions in Multi-agent
Systems. 2011 13th International Conference on
Advanced Communication Technology (ICACT '11).
Phoenix Park, Korea (South): IEEE computer society.
Du, T. C., Li, E. Y. & Chou, D. 2005. Dynamic vehicle
routing for online B2C delivery. Omega, 33, 33-45.
Euzenat, J. & Shvaiko, P. 2007. Ontology matching,
Secaucus, NJ, USA, Springer-Verlag New York, Inc.
FIPA. 2001. FIPA Ontology Service Specification.
[Accessed 28-01-2013].
FIPA. 2002. FIPA Iterated Contract Net Interaction
Protocol Specification. Available: http:// [Accessed 28-01-
Maedche, A., Motik, B., Silva, N. & Volz, R. 2002.
MAFRA - A Mapping Framework for Distributed
Ontologies in the Semantic Web. 13th International
Conference on Knowledge Engineering and
Knowledge Management. Ontologies and the Semantic
Web (EKAW '02). Siguenza, Spain: Springer.
Malucelli, A., Palzer, D. & Oliveira, E. 2006. Ontology-
based Services to help solving the heterogeneity
problem in e-commerce negotiations. Electron.
Commer. Res. Appl., 5, 29-43.
Mei, P. Q., Hong, Z., Cun, C. Y. & Qin, P. X. 2009. An E-
negotiation Model Based on Multi-agent and
Ontology. Proceedings of the 2009 International
Conference on Computational Intelligence and
Natural Computing (CINC '09). Wuhan, China: IEEE
Computer Society.
Nascimento, V., Viamonte, M. J., Canito, A. & Silva, N.
2012. Enhancing ontology alignment recommendation
by exploiting emergent social networks. 2012
IEEE/WIC/ACM International Conference on
Intelligent Agent Technology (WI-IAT '12). Macau,
China: IEEE Computer Society.
O'Brien, T. M. 2009. Google Announces Support for
Microformats and RDFa. Available:
support-for-m.html [Accessed 28-01-2013].
OAA. 2001. OAA [Online]. Available: http:// [Accessed 28-01-2013.
Silva, N., Viamonte, M. J. & Maio, P. 2009. Agent-Based
Electronic Market With Ontology-Services. 2009
IEEE International Conference on e-Business
Engineering (ICEBE '09).
Macau, China: IEEE
Computer Society.
Viamonte, M. J., Nascimento, V., Silva, N. & Maio, P.
2012. AEMOS: An Agent-Based Electronic Market
Simulator With Ontology-Services And Social
Network Support. 24th European Modeling &
Simulation Symposium (Simulation in Industry) (EMSS
'12). Viena, Austria.
Viamonte, M. J., Ramos, C., Rodrigues, F. & Cardoso, J.
2007. ISEM: A Multi-Agent System That Simulates
Competitive Electronic MarKetPlaces. International
Journal of Engineering Intelligent Systems for
Electrical Engineering and Communications: Special
Issue on Decision Support, 15, 191-199.
Viamonte, M. J., Silva, N. & Maio, P. 2011. Agent-Based
Simulation of Electronic Marketplaces With
Ontology-Services. 23rd European Modeling &
Simulation Symposium (Simulation in Industry) (EMSS
'11). Rome, Italy.
Wang, G., Wong, T. N. & Wang, X. H. A Negotiation
Protocol to Support Agent Argumentation and
Ontology Interoperability in MAS-Based Virtual
Enterprises. Information Technology: New
Generations (ITNG), 2010 Seventh International
Conference on, 12-14 April 2010. 448-453.
Wasserman, S. & Faust, K. 1994. Social Network
Analysis: Methods and Applications (Structural
Analysis in the Social Sciences), Cambrige, Cambrige
University Press.