Information Technology Reference
In-Depth Information
delegation (reliance) an agent Ag
1
needs or likes an action of another agent Ag
2
and includes
it in its own plan (see Section 2.9)
.
In adoption (help) an agent Ag
2
acquires and has a goal as (long as) it is the goal of another
agent Ag
1
,thatis,Ag
2
has the goal of performing an action because this action is included
in the plan of Ag
1
. So, also in this case
Ag
2
plays a part in
Ag
1
's plan (sometimes
Ag
1
has no
plan at all but just a need, or a goal) since
Ag
2
is doing something for
Ag
1
.
We consider the action/goal pair
,g)
as the real object of delegation,
21
and we called it
τ
=
(
α
a 'task'. Then by
), to its resulting world state (
g
), or to both.
We introduce an operator of delegation with three parameters:
τ
, we will refer to the action (
α
Delegates
(
Ag
1
Ag
2
τ
)
(7.7)
where
Ag
1
,
Ag
2
are agents and
τ
=
(
α
,g)
is the task. This means that
Ag
1
delegates the task
τ
to
Ag
2
. In analogy with delegation we introduce the corresponding operator for adoption:
Adopts
(
Ag
2
Ag
1
τ
)
(7.8)
This means that
Ag
2
adopts the task
τ
for
Ag
1
:
Ag
2
helps
Ag
1
by caring about
τ
.
Dimensions of Delegation and Adoption
We consider three main dimensions of delegation/adoption:
interaction-based
,
specification-
based
, and
control-based
types of delegation/adoption (Castelfranchi and Falcone, 1998). Let
us analyze these cases more in detail.
Interaction-based types of delegation.
Three general cases may be given:
weak
,
mild
and
strong delegation
. They represent different degrees of strength of the established delegation.
In the following we synthesize (more formal details can be find in Section 2.9) the mental
ingredients of trust in the different delegation actions.
W-Delegates
is the operator for representing
weak delegation.
So the expression:
W-Delegates(Ag
1
Ag
2
τ
)
represents the
necessary
mental ingredients for
Ag
1
trusting
Ag
2
on the task
, shown in Figure 2.11 and resumed in a less formal way in Table 7.1.
We consider in Table 7.1 (
a
,
b
,
c
, and
d
) what the agent
Ag
1
views as a '
Potential for relying
on
' agent
Ag
2
, its
trust
in
Ag
2
; and (
e
and
f
) what
Ag
1
views as the '
Decision to rely on
'
Ag
2
.
τ
21
We assume that
delegating an action necessarily implies delegating some result of that action
(i.e. expecting
some results from
Ag
2
's action and relying on it for obtaining those results). Conversely,
to delegate a goal state
always implies the delegation of at least one action (possibly unknown to Ag
1
) that produces such a goal state as a
result
(even when
Ag
1
asks
Ag
2
to solve a problem, to bring it about that
g
without knowing or specifying the action,
Ag
1
necessarily presupposes that
Ag
2
should and will do some action and relies on this).
Search WWH ::
Custom Search