background image
DigiCULT 31
A
GENT
TECHNOLOGIES
I
ntelligent software agents are an evolving con-
cept of computational entities that should have a
number of properties such as autonomous, flexible
and goal-directed behaviour in open distributed sys-
tems. Other important properties are cooperation in
multi-agent systems and learning, i.e. the agent soft-
ware application needs to be able to learn about
changes in the systems and how to adapt to them
without having to be told.
The classic examples are user-facing applications
such as a personal assistant that helps its owner in var-
ious tasks such as finding specific information on
the Web, the planning of activities such as travel, and
even carrying out reservations and negotiating terms
of payment. These are application scenarios that han-
dle personally delegated responsibility. However, there
are also many possible `behind the scene' applications
such as communications management. Here agents
can be used to increase the local intelligence, thus
relieving networks from heavy communication loads
in the exchange of information between processes.
In fact, using software agents to enhance the cooper-
ation of components of distributed systems is one of
the key application areas heading at more responsive,
intelligent service provision.
Compared with the goals of autonomous, coopera-
tive, flexible and goal-directed behaviour in open sys-
tems, agents today are fairly self-contained, limited to
particular application fields and functionality. Howev-
er, according to the AgentLink Roadmap (2003), the
field of agent technology should mature fairly rap-
idly until 2010. From 2009 or so onwards, it should
see `truly-open and fully-scalable multi-agent sys-
tems, across domains, with agents capable of learning
appropriate communications protocols upon entry to
a system, and with protocols emerging and evolving
through actual agent interactions'.
61
S
EMANTIC
W
EB
S
ERVICES
W
eb Services promise to realise a scenario in
which people and automated mechanisms may
call up and use Internet-based applications on the fly to
perform a variety of tasks and transactions. In fact, Web
Services allow different computer systems running dif-
ferent software applications to communicate with each
other and conduct transactions over the Internet. As
self-describing, automatically discoverable services, they
may become completely decentralised and distribut-
ed over the Internet and accessed by a wide variety of
communication devices.
62
In the next few years Web
Services may also gain considerable momentum beyond
the enterprise level, where they are increasingly used
to eliminate the burden of complex, slow and expen-
sive software integration. Instead, Web Services are used
as the `glue' to achieve reliable, automated integration
between different systems and applications, based on
specifications that are created and approved by respect-
ed bodies such as the World Wide Web Consortium
(W3C) and the Organization for the Advancement of
Structured Information Standards (OASIS).
63
The Semantic Web is envisaged as a Web of
machine-readable information that allows software
agents to carry out tasks for humans automatically. In
particular, it may be one answer to the `information
tsunami'
64
on the Web by creating growing islands of
knowledge representation in an ocean of information
that is poorly described and not semantically interre-
lated to allow for automatic interpretation and effec-
61
M. Luck, P. McBurney
and C. Preist, Agent
Technology: Enabling Next
Generation Computing,
AgentLink, 2003, p. 3.
http://www.agentlink.org/
roadmap/roadmap.pdf
62
Cf. The CBDi Web
Services Roadmap, in
particular: Lawrence Wilkes,
"ROI - The Costs and
Benefits of Web Services",
2003, http://roadmap.
cbdiforum.com/reports/
roi/; Jason Bloomberg,
"The ZapThink Web
services roadmap: It`s going
to be a bumpy ride", 1
July 2002, http://search-
webservices.techtarget.
com/originalContent/
0,289142,sid26_
gci836909,00.html; a lot of
further information may
be found at http://www.
webservices.org; http://
searchwebservices.techtar-
get.com; http://www-130.
ibm.com/developerworks/
webservices/
63
W3C, http://www.
w3.org; OAIS, http://www.
oasis-open.org
64
See School of
Information Management
and Systems, University
of California, Berkeley:
"How Much Information?"
(2003). http://www.sims.
berkeley.edu/research/
projects/how-much-
info-2003/execsum.
htm#summary
DCTHI7_271104.indd 31
06.12.2004 8:37:25 Uhr