Information Technology Reference
In-Depth Information
That is how autonomous robots constructed in realspace became the inspiration
for research on creating similar entities in cyberspace. The fact that cyberspace could
exist as virtual space parallel to realspace and the possibility of existence of certain
entities in cyberspace, e.g. agents, was soon noticed and used in computer science.
If realspace can be the place of action of certain autonomous entities such as robots,
then in cyberspace there can exist and operate agents that have similar characteristic
properties to those which exist in the cyberspace of robots. Finally, the view was
formed that agents created in cyberspace would have similar qualities to those of
robots'—autonomy and capability for observation of the surrounding environment
constituting part of cyberspace.
The first proposals from the 1980s of the last century concerning the use of agents
for the construction of distributed computerized systems can be found in the works of
Fortier [85]. The first projects of different architectures of the agent appeared in the
1990s. One of them is the concept of the agent called agent-0 [161]. That architecture
introduced the term mental state , which served as the place for the knowledge of the
agent and for drawing up its activity directions whereas the agent connects with its
surrounding environment with the use of messages received and sent.
One of the most popular approaches which were drawn up at that time is the BDI
( Belief, Desire, Intention ) architecture. This idea contains some elements borrowed
frompsychology. It provided a basis for effective solutions described, inter alia, in the
work AgentSpeak [154]. Among authors of works significant for the development of
the agent concept were at that time: Maes [121], Rao and Georgeff [154], Demazeau
[87], Calstelfranchi [96], Ferber [81], as well as Jennings andWooldridge [148, 186].
The following years saw the rapid development of research on generating methods
and application possibilities of agent systems.
During that time, in a fewcentres in Poland, research on the concept of an agent and
agent systems was also undertaken. These studies include attempt to define the agent
term, which resulted in the introduction of the first versions of agent architectures
( M-agent architecture [59, 131]), as well as studies on the application of the concept
of an agent in evolution systems [25, 26, 35, 49, 52, 78]. Moreover, studies analysing
the concept of an agent and agent systems according to different points of view were
also conducted [75].
Research on the application of the concept of an agent in different environments
is also interesting and include studies on the application of an agent in the “Middle-
ware” layer in the server environment, and the concepts of agent communication in
the SOA systems [4, 6, 7, 10, 107] in the environment of realspace and cyberspace,
using the concept of an agent-robot [168]. Research on the application of multimodal
logics to the description of agent systems and analysis of the development processes
of agent coalitions [79, 99-101] are also of great interest. The agent system for task
distribution in multiprocessor structures was one of the first projects of a practical
agent system in the country. It was realized within studies carried out between the
1980s and 1990s. The experimental version of the system is presented in works
[30, 33, 38, 133]. The system was supposed to ensure the maximum efficiency of
calculations accomplished due to the multiprocessor structure (e.g., grid or cloud),
Search WWH ::




Custom Search