Agent-based service-oriented architecture for heterogeneous data sources management in ubiquitous enterprise

jumentousmanlyInternet και Εφαρμογές Web

21 Οκτ 2013 (πριν από 3 χρόνια και 9 μήνες)

110 εμφανίσεις

Agent
-
based service
-
oriented architecture for heterogeneous data sources management in ubiquitous enterprise



Agent
-
based service
-
oriented architecture for heterogeneous data
sources management in ubiquitous
enterprise

L.Y. Pang
*

, R
ay
Y. Zhong
, and George Q. Huang

HKU
-
ZIRI Lab for Physical
Internet
,
Department of Industrial and Manufacturing Systems Engineering

The University of Hong Kong

Pokfulam Road, Hong Kong SAR, China

A
BSTRACT

In a ubiquitous manufacturing environment, different devices such as radio frequency identification (RFID)
te
chnology are used to collect real
-
time data. Additionally, data
is

used by different enterprise information
systems

for supporting managerial decision making.
Since data sources from applications and devices are
characterized by multiple types of heterogen
eit
ies

such as communication channels, blinding methods, and
developing environments, the difficulty in managing heterogeneous data sources is greatly increased.
This paper
proposes an innovative Application Information Service (AIS) that serves as a middl
eware for information
exchange in between different applications. The AIS possesses several key contributions. First
ly
, AIS provides a
centralized platform to manage distributed heterogeneous data sources so as to reduce the data duplications,
increase con
sistency, and accuracy. Second
ly
, it combines software agent technologies with service
-
oriented
architecture (SOA) so that services are capable of accomplishing tasks in an autonomous way without human
intervention. Third
ly
, agent
-
based service
-
oriented ar
chitecture paradigm is proposed to cultivate a collaborative
environment to integrate different data sources as well as third party application providers.

1.

I
NTRODUCTION

Enterprise Information Systems (EISs) play an important role in the enterprise.

EISs

provide a platform
for

organization
s

to integrate and coordinate their business processes and activities

such as warehousing, manufacturing,
inventory co
ntrol, etc. The objective
s

of EISs
are

to

increase organization agility, follow the market changes,
imp
rove the relationship between business partners, and enhance the enterprise competitiveness. Since different
EISs provide various functions, different types of EIS may be adopted within a single enterprise

based on the
enterprise situation
.

Also, recent ad
vances in Auto
-
ID technology for supporting operations have increased the
demand for ubiquitous data exchange among Auto
-
ID devices and EISs

[1]
. An enterprise equipped with distributed
and simultaneous information processing capacity is termed as
ubiquitous enterprise.

Under the ubiquitous
enterprise, decision

makers need
comprehensive

information for high
-
level or managerial decis
ion making
.

This paper is motivated by real
-
life problems from
a
collaborating company which encounters

problems in
ma
naging different EISs.
The
collaborating company

is a
public warehouse

providing logistics and value added
services for its manufacturing partners. The company has a small production line for repacking manufacturers’
finished product according to the custo
mer requirements. The packaged product will then consolidate and ship to
customers.

The company is facing some problems in managing different EISs
.
Firstly,
the
company adopted

several
EISs to support their daily operations. However, operating
data

are

inp
utted and stored in various EISs separately
and independently. Thus, data

is

usually

asynchronous, inconsistence, and inaccuracy
in
between EISs. Secondly,
transactions and data exchange between the company and its customers are

handled

by employees

manual
ly
. Human
errors at this time are unavoidable and require a lot of time and efforts to ensure the data
correctness
. Also,
the

format and structure

of information
from different customers are
various
.
The company

pay
s

a lot of
human
resources in order

to

ensure data consistence between EISs
. This increases the operating costs and the responsive
time
and hence

affect
ing

the
company

image
. Third
ly
, the company uses a lot of Auto
-
ID devices such as RFID,
barcode devices to improve the

performance of
inventor
y

management. However, the

outputting

format and
structure
of
Auto
-
ID

devices

are highly irregular.
The company

found

that

it is
difficult

for them to integrate
real
-
time data and existing static
information to support decision

making.





*

Corresponding author: E
-
mail: h0664249@hku.hk


Flexible Automation and Intelligent Manufacturing, FAIM2013



To address the abov
e problems,

significant research progresses have been made

in enterprise information
integration. The objective of enterprise information integration is to provide
a

uniform access to various data sources
[2]
.
Sujansky summarized four significant represent
ational heterogeneity differences for aggregating
data;

they are
structural differences, naming differences, semantic differences and content differences
[3]
. A lot of methods have
been proposed for data integration. Halevy summarized and categorized data integration in 6 directions
[4]
. 1)
Using

mediated schema to map multiple data sets to one schema. 2)
Developing

an adaptive query language to query
multiple

data sources.

3)
Using

XML to integrate data. 4)
Developing

models to provide
algebra

for manipulating
and mapping different data sources. 5)
Using

peer
-
to
-
peer technology to share data or files. 6)
Using

artificial
intelligence technologies
to semi
-
autom
atically generates semantic mapping for data integration.

A

lot of
frameworks have been proposed
to
integrat
e

and merg
e

enterprise information system systematically
[5
-
7]
. Artificial
intelligence technologies such as software agent technology have been wid
ely adopted in information system
integration
[8]
. Software a
gent

technology is an important branch of artificial intelligence and has been widely
accepted and adopted in different areas such as supply chain, manufacturing, and product design for it

charac
teristics
of autonomy, interoperability, flexibility, re
-
configurability and scalability
[9
-
11]
.
Agent
-
based approach
es have

played an important role to achieve outstanding performance with agility in various kinds of applications such as
resources managem
ent

[12]
, planning and scheduling

[13]
, and information processing

[14]
.

Foundation for
Intelligent Physical Agents (FIPA) provide
s

an overview and guide to design and develop heterogeneous interact

agents and agent
-
based systems

[15]
.

Java Agent Developme
nt (JADE) framework that compliant to FIPA
specifications is a common framework for agent applications under JAVA environment
[16]
.
The

recent movements
of enterprise information integration are focusing

on the

integration of web service technology and artificial
intelligence technologies
[17
-
19]
. Web services technology provides a higher
-
level interoperability for leveraging
business activities across the web either within an enterprise or among collaborating e
nterprise.
Web Services
technology and Service
-
Oriented Architecture

(SOA)

have provided a new and excellent solution to the data
integration among heterogeneous and distributed systems. Agents and
web services

are integrated together to
provide a more fle
xible and dynamic solution

[20]
.

However, the knowledge requirements and technology threshold
of implementing these technologies are very high. Small and middle enterprise may not have enough knowledge and
expertise in developing
these complicated systems
.

Therefore, several research problems are addressed in this paper. Firstly, how to

integrate and manage the
heterogeneous

data sources
, such as EISs and Auto
-
ID devices, in ubiquitous
enterprise

such

that they are

usable by
non
-
technical end
-
users? Secondl
y, how to wrap and integrate the multi
-
types data sources and export the result to a
common format so that data can be easily shared? Thirdly,
h
ow to manage the agents and its services through web
services
technologies

in order

to make them be easily regis
tered, discovered, and invoked by users, decision support
systems or other applications?

The aim of this research is to design and develop an

agent
-
based

application
information service (AIS)
, which adopted SOA
, for managing heterogeneous data sources in u
biquitous enterprises.
The objective
s

of this paper
are

firstly
,

it is

to develop a framework of AIS platform
that users can

integrate and
manage heterogeneous data sources. Second
ly
,

it is

to develop a mechanism

that

wrap
s

heterogeneous data sources
so th
at

international standards such as ISA
-
95

can be easily applied

to standardize data enquiry, format and
structure. Third
ly
,

it is
to

integrate agents with web services

that

promote

services intelligence and collaboration.
Also, it enhances agents so that they are

easily

be

registered, discovered, and invoked.

This research adopts and
de
velops three important concepts. Firstly, intelligence agent technology is used as a mediator that

provides
translation of queries and data between heterogeneous data sources. Secondly, web service technology is used for
promoting agent operation
s

to agent services

that enhance the software agent characteristics.

Thirdly,

a
n agent
-
based
service oriente
d architecture that adopts SOA in designing the agent architecture for efficient integration of different
AIS agents.

The rest of the paper is organized as follows. Section 2

is the overview of the agent
-
based data source
management. Section 3 is the frame
work of the AIS. Section 4 is the design of agent
-
based SOA
. Conclusion and
future works are given in section
5

to summarize this research.


2.

O
VERVIEW OF
A
GENT
-
B
ASED
D
ATASOURCES
M
ANAGEMENT

Under the ubiquitous enterprise,
several EISs are usually used in

an enterprise to support their daily operations.
Information stores in different EISs separately. If there is no systematic mechanism for data management,
information become asy
nchronous, inconsistence, and disperse in different EISs. Therefore,

an agent
-
based AIS
platform is proposed to assist users in managing heterogeneous data sources.
The overview of Agent
-
Based
Applica
tion Information Service is shown in
Figure
1
. AIS acts as a middleman in between data consumers and data
providers to

provide
several

services.
The first service is

to

manag
e

different data sources within the enterprise.
The
second service is

to

integrat
e

data from different data
sources in order to provide comprehensive information for
decision

making.
The third service is

to

ensur
e

data is

consisten
t
, non
-
duplicated and
accura
te

in different data
Agent
-
based service
-
oriented architecture for heterogeneous data sources management in ubiquitous enterprise



sources. Finally,

it is to

provid
e

a set of visual tools for general users to (re)co
nfigure and manage the AIS platform
and various data sources.


An example is selected to illustrate the situation
.

Work
-
in
-
progress (WIP) is

important

information for decision

making within the manufacturing enterprise.

WIP information is
used

t
o 1) calcu
late total inventory value, 2) plan
the production schedule and 3) monitor the production progress. However,
WIP
is comprehensive information
which

usually consists of
four

set
s

of basic information
. It
includes

item

information from ERP syste
m, item
quantity and location from warehouse management system (WMS), scheduling and planning information from
MES, and real
-
time production such as WIP quantity and location information from shop
-
floor. T
he data consumer
such as decision support systems sends the

data request token to the AIS
Platform. The token indicates what
information is needed to retrieve.
This

information is defined in the standard data structure template repository
which act
s

as a library for different information sets. After the Manager Ag
ent receive
s

the request, it will find and
invokes different data sources agent which responsible for accessing and collecting request data from different data
sources. Final
ly, the Manager Agent integrates

all data piece
s

and send
s

the result back to the
data consumers.

Figure
1
: Ove
rview

of Agent
-
Based

Application

Information Service

3.

F
RAMEWORK OF
A
GENT
-
BASED
A
PPLICATION
I
NFORMATION
S
ERVICE

Different data sources are driven by different communication methods such as class
libraries, communication
protocols,

and

invoking methods. In this regard, professional programming knowledge and expertise are obviously
needed. Moreover,
a tiny change

of business processes and requirements may need to redesign and recoding the
existing a
pplicati
ons in order to collect right information. It is, therefore,
necessary

to configure and manage the
heterogeneous data sources following a uniform and flexible model
to reduce the technical difficulties
. For the
above purpose, the proposed AIS is to

integrate the concepts of software agent and web services.

Figure
2

shows the
framework of agent
-
based AIS. It aims to provide a framework for agent management services, and a platform for
agent message transport. The purpose of AIS is to help data consumers
collect data from different data providers and
Flexible Automation and Intelligent Manufacturing, FAIM2013



integrate those data to meaningful information with standard data structure. Inside the AIS, software agent
technology is adopted for managing and accessing different data sources.

It is used for wrapping vari
ous
heterogeneous data sources to form a uniform architecture. The most innovative features of AIS is to provide a
standard data structure template library for users to customize the output data format and wrap heterogeneous data
sources to a uniform inter
face for data capturing.

The design of agent
application

in AIS is compliant

with

the FIPA
specification
.

The agent
-
based AIS includes four main modules.

Figure
2
: Framework of Agent
-
Based AIS

1.

AIS Agents

This module is the coll
ection of agents and their related services. There are three types of agents. They are
Manager Agent, Data Source Agents (DS Agents), and Wrapper Agents.

The Manager Agent is the mediator
between data consumers and other agents. The AIS manager has three p
urposes.
The first purpose is to handle

requests from data consumer.

The second purpose is to communicate

different DS agents to retrieve data from data
providers.
The third purpose is

to convert

different data sets to standard data structure and
return

th
e result to data
consumers.

The purpose of DS agents is to access different data sources that are wrapped by Wrapper Agents. The
purpose of Wrapper Agent is to wrap non agent
-
based data sources to a uniform
interface

and join the agent
-
based
AIS platform.

2.

AIS Management Tools

The main purpose of this module is to provide visual tools and APIs

to manage and configure the AIS platform
and
AIS

agents. There are three tools in this module. The first one is Data Structure Template Management tool.
This
tool a
llow
s

uses manage the standard data structure template repository. This repository is a library for
outputting data format. Data retrieve from different data source may contain different data structure, table name,
column name, etc. Users can use this tool

to standardize the data structure that they want to retrieve. The second tool
is Data Conversion Rules Management. After users establish the standard data structure, users need to use this tool
to create the converting rules or the matching table for tran
slating data from the original data format to the requiring
data structure.

The third tool is Data Source Agent Management which enables users to manage AIS Agents
Agent
-
based service
-
oriented architecture for heterogeneous data sources management in ubiquitous enterprise



manually and checks AIS Agents status.
The AIS management tool

exerts supervisory control ov
er the

whole

AIS
platform.

3.

AIS
-
UDDI

This module is an agent directory facilitator that provides a yellow pages directory services to agents and their
services. An AIS
-
UDDI platform, which is
XML
-
based

registry, is developed and used for
agent services
describing
and discovering through internet
.

The detailed structure will be introduced in the next section.

4.

Message Transport Service

This service is used for delivering message between agents within the agent platform and to agents that on others
agent

platform. The communication between agents is through agent message which consists of a message envelope
and a message body. Agent messages will be routed and h
andled

between agents according to the specific transport
requirements such as transport protoc
ol.

4.

A
RCHITECTURE OF
AIS

A
GENTS AND
S
ERVICES
D
IRECTORY

As discussed above, AIS agents are exposed as web services form. The SOA is adopted to design the AIS
agent
directory facilitator
. The SOA cultivates a loosely coupled environment that supports dyna
mic binding agents or
services during run
-
time, and standardize
s

the different agents’ communication pr
otocol. The
Figure
3

shows
the
architecture of A
IS
Agents and S
ervices
D
irectory. An AIS
-
UDDI platform, which is
XML
-
based

registry, is
developed and used for
agent services describing and discovering through internet
. In the implementation level,
web
services, WSDL, and SOAP provide such capacities as
discovery, deployment and communication to realize such
architecture. Web services are commonly adopted to develop internet accessible applications. The

services, which
are provided by AIS Agents,

are

compose
d

and encapsulate
d

in web service format in orde
r to standardize the
accessing protocol and invoking methods. WSDL describe
s

the operations correspond to the action supported by the
agents. SOAP is a

platform
-
independent

communication protocol for exchanging XML
-
based message via HTTP
protocol.

This rel
ationship of web services, AIS
-
UDDI, WSDL,

and SOAP cultivate

a
SOA

for AIS

A
gents

and
Services Directory.

A complete process of using AIS
-
UDDI involves three main phases: service registration, searching, and binding.
In the registration phase, after depl
oy agents in web services format, the information of agents and its related
services should be registered to the AIS
-
UDDI to indicate its availability, capability, location, as well as the
interfacial description. Then, the registered agents and their prov
ide services can be searched by other agents. In the
se
arch phase, other agents or sub
-
agents can discover and obtain other agents information. The AIS
-
UDDI provides
two ways for web services discovery, programmatically, or through the service searching explorer GUI. The former
means the discovery could be triggered by SOAP message sent from

the program of a remote requestor, while the
latter means the requestor could use the service searching explorer to manually, perform the service discovery. The
searching result will indicate the following information: 1) the uniform resource identifier (
URL) of the agent
location / provider. 2) the remote service(s) / method(s) name for invoking. 3) the input and output parameters data
types. In the binding phase, once the request prepares proper input data for intended the agent service, it starts
bindin
g the service URL and invoking the service method. The communication between requestor and receiver is via
SOAP request message. The receiver agent decodes the message, executes action and returns the XML
-
based result.
The request agent parses the result f
or its own process.

Figure
3
:
the Architecture of AIS

Agents and Services
Directory

Flexible Automation and Intelligent Manufacturing, FAIM2013



5

A
GENT
-
BASED
A
PPLICATION
I
NFORMATION
S
ERVICE
D
ATA
P
ROCESSING
P
ROCESS

In this section,
a

product information requesting
scenario is envisaged to
demonstrate the agent
-
based AIS data
processing

procedure
.

Product information is
important

and critical information required for upper level decision
making in manufacturing enterprises. Enterprises require product information in calculating the inventory

value,
costing, production planning and scheduling. Production information in manufacturing enterprise commonly
consists of four sets of information. They are basic production information form ERP, product quantity and location
from warehouse management s
ystem (WMS),
product BOM list and production progress from MES, and WIP
information from shop
-
floor. Detailed procedure of the aforementioned processes is shown in
Figure
4
.

Figure
4
: AIS Agents
Data

Processing

Procedure

From the perspective of processes, the AIS Agents consist of three different levels of agents, from the top to
down, are Manager Agent, Data Source Agents (DS
Agents), and Wrapper Agents. The higher levels of agents
provide more complex services and the lower level of agents provides more basic and primitive services. Different
levels of agents have different objectives to achieve. The AIS Manager is mediator be
tween data consumer and data
source agents. The purpose of

the

Manager Agent is to coordinate the sub level data source agents and integrate data
which retrieves from DS Agents to a standard data structure. The purpose of DS Agents provide basic services f
or
communicating different types of data sources such as database, devices, or application systems because the
parameters input are very similar. The purpose of Wrapper Agents act as an adaptor function to encapsulate
non
-
agent data sources and allow DS ag
ents to access them as an agents via a standard communication protocol.
Also, the DS Wrappers provide two functions. The first function is to integrate the drivers, class library, and
communicate protocol

that related to
the data sources. The another funct
ion is to convert the agent message to query
language such as SQL for database, SOAP message for web services, and other dedicated invoking methods for
Auto
-
ID devices.

The steps of Data Processing Procedure

are described

as follows. First, the data consum
ers such as
decision support systems will send a request token to Manager Agent. The request token is XML based and
information includes three types of information, the

required

data structure template to retrieve, the target data
providers, and filtering
parameters.
Second, the
M
anager

Agent

receive
s

token and translate
s

the token. Then, the
correspondence data template is loaded from the native data template repository. Forth, according to the

data

Agent
-
based service
-
oriented architecture for heterogeneous data sources management in ubiquitous enterprise



structure template, the AIS manager search
es

and invoke
s

different type of DS Agents to collect information from
different data sources. Fifth,
the DS Agent parses the message and discovery
suitable Wrapper

Agents to access

the

target data source
. Sixth, the Wrapper

Agent
translates

the message
to corresponding
query language with proper
condition parameters. Then, the wrapper

agent

accesses the data sources with suitable class library, protocols and
methods to connect and retrieve information from the data source. The result finally transfers back to the
M
anager

Agent
. Eighth, after the
M
anager

Agent collects all required information from

DS agents
, all piece of information
will be integrated together
based on

the

specifications of

data structure template.

Finally, the result send
s

back to the
data consumer for
f
urther processing

or decision

making
.

6.

C
ONCLUSIONS AND
F
UTURE
W
ORK

In a ubiquitous manufacturing,
data
is

usually

stored in heterogeneous data sources. Since each data source has
its

own data structure and format, unmanaged information easily becomes inconsistent, duplicated, and inaccurate.
There are an increasing number of data sources with need for uniform query interfaces to access and manage
distributed data across diverse source
. This paper has proposed an agent
-
based application information service to
manage heterogeneous data sources in ubiquitous manufacturing. A standard data structure library is established for
storing the data output structure template. AIS Agents are devel
oped and used for accessing heterogeneous data
sources and integrating data from different data sources according to the selected data structure template. The united
data format enable data interchange between applications. The use of agents enhances the f
lexibility and scalability
of the AIS.

The agent
-
based AIS

demonstrates several key contributions.

First
ly
, AIS provides a centralized platform to
manage distributed heterogeneous data sources so as to reduce the data duplications, increase

data

consistenc
y and
accuracy.
Also
, the

use of standard data structure template library

is

to enable users self
-
customized the output data
structure.
S
tandard schema can be added to the data template library easily. The use of standard

schema

facilitates
the information

interchange between applications as well as enterprises
.

Second
ly
, it combines software agent
technologies with

SOA so that services are capable of accomplishin
g tasks in an autonomous way without human
intervention. Services provided by AIS agents are wr
app
ed by web services that provide

a standard communication
interface for communication among agents.
This can enhance the agents’ communication and cooperatively
in

cross
-
platform and cross
-
enterprise environment.

Third
ly
, the design and development of AI
S
-
UDDI provides a
platform to register, publish, search and bind the AIS agents and their related services. The implementation of
AIS
-
UDDI realizes the SOA and cultivates a collaborative environment to integrate different data sources as well as
third
-
part
y application
s
.

Operation

p
rocesses are created by linking up different individual services. Therefore,
IT
-
engineers are easy to reconfigure, rearrange, and reuse different services. This maximizes system integrity and
scalability.

The current work will be

further extended in the future research form several aspects. First
ly,

the data source
agents and wrapper agents will be extended to support more type of data sources through accommodating more
drivers and API library. Second
ly
, the standard template libr
ary will be extended to support more international
standard in logistic and manufacturing domain. Therefore, more efforts are necessary to better understand these
standards for better exploitation. Finally, ontology will be implemented in the agents to hel
p for integrating data
between data sources. Data syntactic, schematic and semantic heterogeneities are main barriers affect the data
integration process. The next step is to implement ontology methods to solve this type of problem.



Flexible Automation and Intelligent Manufacturing, FAIM2013



A
CKNOWLEDGEMENTS

Authors are grateful to the Zhejiang Provincial, Hangzhou Municipal and Lin’an City governments for partial
financial supports. HKSAR RGC GRF HKU 712112E, Guangdong Modern Information Service Fund 2009
(GDIID2009IS048), 2010 Guangdong Department of Science

and Technology Funding (2010B050100023), and
International Collaborative Project of Guangdong High Education Institution (gjhz1005).

R
EFERENCE

[1]

Mohsen, A., RFID: an enabler of supply chain operations. Supply Chain Management: An International Journal,
2007.
12(4): p. 249
-
257.

[2]

Halevy, A.Y., et al., Enterprise information integration: successes, challenges and controversies, in Proceedings of the
2005 ACM SIGMOD international conference on Management of data2005, ACM: Baltimore, Maryland. p. 778
-
787.

[
3
]

Sujansky, W., Heterogeneous database integration in biomedicine. Comput. Biomed. Res., 2001. 34(4): p. 285
-
298.

[4]

Halevy, A., A. Rajaraman, and J. Ordille. Data integration: the teenage years. in VLDB'2006: Proceedings of the 32nd
international confe
rence on Very large data bases. 2006. VLDB Endowment.

[
5
]

Ronald, E.G., A framework to review the information integration of the enterprise. International Journal of Production
Research, 2004. 42(6): p. 1147
-
1166.

[
6
]

Lenzerini, M., Data integration: a the
oretical perspective, in Proceedings of the twenty
-
first ACM
SIGMOD
-
SIGACT
-
SIGART symposium on Principles of database systems2002, ACM: Madison, Wisconsin. p. 233
-
246.

[
7
]

Woelk, D., et al. Using Carnot for enterprise information integration. in Parallel a
nd Distributed Information Systems,
1993., Proceedings of the Second International Conference on. 1993.

[
8
]

Gorton, I. and A. Liu. Architectures and technologies for enterprise application integration. in Software Engineering,
2004. ICSE 2004. Proceedings.

26th International Conference on. 2004.

[
9
]

Sikora, R. and M.J. Shaw, A Multi
-
Agent Framework for the Coordination and Integration of Information Systems.
Management Science, 1998. 44(11): p. S65
-
S78.

[
10
]

Gek Woo, T., C.C. Hayes, and M. Shaw, An intellig
ent
-
agent framework for concurrent product design and planning.
Engineering Management, IEEE Transactions on, 1996. 43(3): p. 297
-
306.

[11]

Jiao, J., X. You, and A. Kumar, An agent
-
based framework for collaborative negotiation in the global manufacturing
s
upply chain network. Robotics and Computer
-
Integrated Manufacturing, 2006. 22(3): p. 239
-
255.

[
12
]

Bastos, R.M., F.M. de Oliveira, and J.P.M. de Oliveira, Autonomic computing approach for resource allocation. Expert
Systems with Applications, 2005. 28(1):
p. 9
-
19.

[
13
]

Weiming, S., W. Lihui, and H. Qi, Agent
-
based distributed manufacturing process planning and scheduling: a
state
-
of
-
the
-
art survey. Systems, Man, and Cybernetics, Part C: Applications and Reviews, IEEE Transactions on, 2006.
36(4): p. 563
-
577
.

[
14
]

R. J. Bayardo, J., et al., InfoSleuth: agent
-
based semantic integration of information in open and dynamic environments.
SIGMOD Rec., 1997. 26(2): p. 195
-
206.

[
15
]

http://www.fipa.org
.

(accessed May, 2012).

[1
6
]

B
ellifemine, F. and G. Rimassa, Developing multi
-
agent systems with a FIPA
-
compliant agent framework. Softw. Pract.
Exper., 2001. 31(2): p. 103
-
128.

[
17
]

Huhns, M.N., Agents as Web services. Internet Computing, IEEE, 2002. 6(4): p. 93
-
95.

[
18
]

Dale, J., et
al. Implementing agent
-
based web services. in AAMAS 2003 Workshop on Challenges in Open Agent
Environments, Melbourne Australia. 2003. Citeseer.

[
19
]

Huhns, M., Software Agents: The Future of Web Services, in Agent Technologies, Infrastructures, Tools, and

Applications
for E
-
Services, J. Carbonell, et al., Editors. 2003, Springer Berlin Heidelberg. p. 1
-
18.

[
20
]

Brazier, F.M.T., et al., Agents and Service
-
Oriented Computing for Autonomic Computing: A Research Agenda. Internet
Computin
g, IEEE, 2009. 13(3):

p. 82
-
87.