Introduction to the Subject of an Agent in Computer Science - A Perspective on Agent Systems - page 4

which is connected with possible evenly distribution of tasks among processors of
structure, under conditions when new demands appear dynamically, and not evenly in
the particular processors. In order to realize the distribution of tasks, the
application of an agent was proposed (which at that time was not known as the
agent). In the work [33], a task for the agent (referred to as “traveling salesmen”)
was defined, as well as the general method of its accomplishment.
Despite the research on the concept of an agent and agent systems conducted
in many global research centers, some of the problems have not found satisfying
solutions. It even applies to the terms connected with the definition or the basic
properties of an agent.
One of the most frequently used definitions of an agent was formed byWooldridge
[184], which can be briefly presented as follows: “An agent is a computer system
that is situated in some environment , and that is capable of autonomous action in
this environment in order to meet its design objectives”.
According to this definition it can be concluded that the term agentautonomy
provides a basis for defining the agent.
Although the definition is succinct, it refers to certain, not clearly defined terms.
The crucial term autonomous is defined intuitively, mainly with the use of examples
presented and discussed in relevant works. The fact that the agent should have the
possibility of interaction with its surrounding environment through observation and
actions realized in the environment is emphasized in these works.
Having analysed this definition, it may be concluded that it is of intuitive character.
In the course of research development on the term agent and agent systems that kind
of intuitive approach appears to be insufficient, especially when we try to apply more
precise methods to examine these systems.
It seems that the definition of the term agent causes difficulties because the goal
of implementing (and even inventing) this term into computer science has not been
clearly defined from the very beginning. Therefore, it seems reasonable, for this goal,
to propose the concept of decomposition of too complex (to realize) algorithms and
farther programs.
However the aim of the decomposition is not so much a matter of accelerating
the running of these programs as of making the very formation process (analysis,
designing, programming) of these algorithms, programs or computer systems easier,
simpler and faster.
It can be stated that the term agent is one of the results of searching for different
(new) ways of these decomposition.
This monograph presents the concept of the agent and agent systems from a for-
mal approach to the examples of practical applications. Starting with a certain formal
definition of an algorithm (using such terms as a set and partial function), the goal of
introducing the agent was defined as a certain paradigm of designing and program-
ming computer systems, specifying its basic properties at the same time (Chap. 2 ) .
In order to form the principles of construction of autonomous agents, a model of
the agent was introduced (Chap. 3 ) . On the one hand, it presents functionalities a sin-
gle agent can be equipped with, on the other hand, it indicates the principles of coop-
eration of groups of agents within the system defined as MAS multi-agent system.