Agent Versus Decomposition of an Algorithm - A Perspective on Agent Systems - page 19

We may improve this relationship of algorithms and so that the algorithms Al g 1
and Al g 2 become autonomous (mutually). It can be realized in two steps, which will
be described further in later chapters.

2.5.1 Step One—The Introduction
of the Environment Concept

The first step is to extract a certain part of variables as the global variables, which
involves the introduction of an environment, described by these global variables
representing the state of that environment.
The inspiration for the introduction of the environment concept is derived from the
observation of mobile robots' action. Their environment plays a crucial role in robots'
action and it has an influence on the design of their algorithms of action, in a sense, the
concept of the environment has accompanied the development of computer science
sine operating systems were introduced. It was operating systems that constituted
and constitute the environment for the realization of programs (algorithms) of the
user. The introduction of the environment concept is realized as follows:

• We choose certain variables and consider them as the parameters describing the

state of the environment. These variables (a set of these variables denoted by X 0 )
are available to each partial algorithm (see Alg1 and Alg2 in Fig. 2.9 ).

•

Other variables are grouped in such a way that each group (of variables) may be
considered as variables describing the state (also referred to as internal state) of
individual partial algorithm ( X 1 , X 2 in Fig. 2.9 ).
We may now consider the component algorithms Al g 1 = ( X 0 × X 1 × X 2 , f 1 )

and Al g 2 = ( cooperating (described in the next subchapter) in the

environment X 0 . The introduction of the environment, the state which is available to
both algorithms and can be changed by these algorithms, may create the impression
that the algorithms to some extent become mutually dependent and lose autonomy
towards each other. Nevertheless, new possibilities appear. Such dependence through
the environment allows cooperation between the algorithms and a replacement of one
algorithm for another, an equivalent one.

X 0 × X 1 × X 2 , f 2 )

Fig. 2.9 Schema of
decomposition of the set X
which results in the partial
algorithms Al g 1 and Al g 2
and the state of the
environment ( X 0 )