[Modeling] Doubt about Roles and Joint Actions
Wed, 1 Oct 2003 19:27:08 +0200
GAIA describes the behaviour of agents in terms of Protocols and
Activities: A protocol is a high level interaction amongst a set of
roles that encapsulate a coordinated task between them, and an activity
is a computation performed by a role without interaction with others.
In our approach, we also work with such multi-role interactions in order
to describe the behaviour of agents.
A role represents a partial view of the behaviour of an agent (playing
several roles) which offers a certain interface to collaborate with
others. The problem is that if we look at a finer granularity, when a
role participates in several GAIA protocols (or an mRIs in our
approach), should it be recursively composed of a set of roles?