Information Technology Reference
In-Depth Information
However, choosing a prototypical organization approach for a whole network may
not be sufficient. In fact, heterogenous relationships may be required between agents in
different parts of the supply network. Moreover, predetermining agent interaction pat-
terns will necessarily lead to a compromise between efficient operation and adaptive be-
havior: For example, negotiation based interaction paradigms are highly adaptive when
it comes to changing behavior of participating agents (as they allow for determining the
best result under any given conditions). Nevertheless, they lead to a large overhead of
communication and computation effort as every interaction task involves all possible
participants among the agents.
In order to confine the interaction effort [18], a MAS can be subdivided into teams
of agents with similar properties or joint objectives [21,22]. Team building and joint
action among autonomous agents for distributed problem solving includes determina-
tion of potentials for cooperative acts, formation of teams, distributed planning, and the
actual processing of plans [22]. In the logistics domain, team formation methods have
shown benefits in terms of increased resource utilization efficiency while reducing the
communication effort of agents performing similar tasks [17,19].
However, clustering agents in teams usually focuses on short term behavior and
tasks, rather than on middle and long term structures in agent interaction. Furthermore,
team formation processes rely on the exchange of information about agent properties
and goals among the potential team members. Hence, they assume any participating
agents to behave benevolently, i.e., to be trustworthy. In an open system, however,
agents may be confronted with deceitfully behaving participants [13] or others, sim-
ply not willing to share information.
Thus, potential interaction partners in open MAS cannot be assumed a priori to ex-
hibit particular behavioral characteristics. In fact, they appear as
black boxes
and there-
fore must be observed by the other agents or the system designer in order to determine
their characteristics during runtime of the system. Based on such observations, a struc-
turing approach for MAS has been proposed, using explicit modeling of
expectations
concerning communication flows [1,14]. This approach, which is inspired by the soci-
ological theory of communication systems [11], establishes a notion of communicative
agent behavior that is reflected by the modeled expectations.
Feeding those expectations back into the decision-making process of interacting
agents offers a promising foundation for self-structuring MAS, as they reflect other
agents' characteristics inferred from their observable behavior. Customer demands, for
instance, can be observed from the incoming orders on the supplier's side. The supplier
can establish expectations regarding the customers' behavior and subsequently adapt
his own behavior with regard to these expectations. Hence, the system as a whole is
enabled to adapt to implicit characteristics and external impacts by the agents refining
their communication patterns in terms of business relationships, i.e., the system orga-
nizes itself.
To summarize, agent coordination refers to communication processes between these
agents. Prototypical coordination mechanisms lead to a compromise between opera-
tional efficiency and flexibility while dynamic team formation requires additional
behavioral assumptions to overcome this problem. However, the systems-theoretical
perspective of expectations structuring agent interaction (rather than assumptions and
