There may be a one-to-one correspondence between roles and agent types, although a number of closely related roles can be
mapped into in the same agent class for the purposes of convenience or efficiency.
Such a correspondence naturally derives from the methodological process that was adopted in which, especially during
the identification of the organizational structure, the concept of role has implicitly assumed ever greater
concreteness.
However, a designer can choose to package a number of closely related and strongly interacting roles in the same agent
class for the purposes of convenience, after ensuring that (i) the organizational efficiency is not affected, (ii)
bounded rationality problems do not emerge, and (iii) this does not violate organizational rules.
There is obviously a trade-off between the coherence of an agent class (how easily its functionality can be
understood), the efficiency considerations, and the need to minimize mismatches with the real-world organization that
the system intends to support.
See Developing multiagent systems: The Gaia methodology (p. 361-362) for more
information about this task.
|