Page 277 - Socially Intelligent Agents Creating Relationships with Computers and Robots

260 Socially Intelligent Agents

ology underpinning their development, and so they are produced in an ad hoc
fashion.
Another fundamental aspect is to opt for either a micro (agent-centered)
view or a macro (organisation-centered) view of MAS. Although early work in
DAI identified the advantages of organisational structuring as one of the main
issues in order to cope with the complexity inherent to designing DAI sys-
tems (f.i.[2]) MAS research has traditionally kept an individualistic character,
evolving patterned on a strong agent-centered flavour. And yet, there is an in-
creasing interest in incorporating organisational concepts into MAS as well as
in shifting from agent-centered to organisation-centered designs [1, 5, 7] that
consider the organisation as a first-class citizen. Nonetheless, in general the
introduction of social concepts into multi-agent systems has been undertaken
in a rather informal way.
In this paper we adopt a macro perspective in order to effectively construct
open multi-agent systems. Thus we argue on the need for deploying norma-
tive environments similar to those provided by human institutions following
the pioneering work in [5]. Institutions [6] represent the rules of the game in a
society, including any (formal or informal) form of constraint that human be-
ings devise to shape human interaction. They are the framework within which
human interaction takes place, defining what individuals are forbidden and
permitted and under what conditions. Furthermore, institutions are responsible
for ascertaining violations and the severity of the punishment to be enacted.
We uphold that open multi-agent systems can be successfully designed and
implemented as institutionalised agent organisations (henceforth electronic in-
stitutions).
In Section 2 we present a case study of human institution in order to subse-
quently identify its components in Section 3. Next, in Section 4 we describe
the two types of agents on which we found a computational model of elec-
tronic institution which successfully served to realise an actual agent-mediated
electronic auction house. Finally, Section 5 contains some conclusions.

2. The Fish Market. An Actual-world Human Institution

As a starting point for the study of institutions we choose the fish market as
a paradigm of traditional human institutions. The actual fish market can be de-
scribed as a place where several scenes take place simultaneously, at different
places, but with some causal continuity. Each scene involves various agents
who at that moment perform well-defined functions. These agents are subject
to the accepted market conventions, but they also have to adapt to whatever
has happened and is happening at the auction house at that time. The principal
scene is the auction itself, in which buyers bid for boxes of fish that are pre-
sented by an auctioneer who calls prices in descending order —the downward