Information Technology Reference
In-Depth Information
components of the motive state and the belief of the agent to compute the motivational
values of desires.
For the computation of the motivational value of a desire we have to combine the
coupling strengths with the level weights for each motive coupling. Based on these
basic motivations we need to consider the interaction of motivations by combining the
basic motivations for the same desire from different motive couplings to determine the
resulting motivational value.
For the computation of the basic motivations we introduce a function that combines
coupling strengths with the level weights for a motive coupling. For this function from
[0
,
1]
×
[
−
1
,
1]
to
[
−
1
,
1]
we demand the properties associativity, commutativity, mono-
tony and having
as its neutral element. Hence, this function should be some kind of a
t-norm (appropriately extended to the interval
1
[
−
1
,
1]
) and we picked the product as a
sufficient candidate for the following definition.
Definition 7.
Given a motive coupling
mc
, its coupling strength
cs
(
mc
)
and level
weight
w
(
mc
)
we define a function
β
:[0
,
1]
×
[
−
1
,
1]
→
[
−
1
,
1]
representing a
basic
motivation
as:
β
(
w
(
mc
)
,cs
(
mc
)) =
w
(
mc
)
·
cs
(
mc
)
.
Assuming positive values of the coupling strength, the influence of the level weight
and the coupling strength is symmetric, for a level weight or coupling strength of
0
the resulting basic motivation will be
the basic motivation
is limited by the coupling strength and vice versa. For values of the level weight and
coupling strength other than
0
. For a level weight of
1
0
and
1
the basic motivation is smaller than both, e. g.,
0
. For negative values of the coupling strength the behavior of the absolute
value is the same but the resulting basic motivation is negative, thus acting against the
realization of the associated desire.
For each desire there might exist several motive couplings and hence a set of basic
motivations results for each desire. In order to determine the motivation value of a desire
we combine the basic motivations for it. This combination function has to account for
the nature of interaction of different motives for the same desire. In order to realize an
adequate combination method, we base our approach for the combination of basic mo-
tivations on the parallel combination initially used for certainty factors in the MYCIN
expert system [2] and using the commutative and associative aggregation function
f
:
(
−
1
.
5
·
0
.
5=0
.
25
,
1)
×
(
−
1
,
1)
→
(
−
1
,
1)
defined as
⎧
⎨
x
+
y
−
x
·
y,
if
x, y >
0
f
(
x, y
)=
x
+
y
+
x
·
y,
if
x, y <
0
⎩
x
+
y
1
−
min
{|x|,|y|}
,
else
we define the motivation value
μ
(
D
)
of a desire
D
as follows.
Definition 8.
Let
D
∈D
,let
mc
1
,...,mc
l
be all motive couplings with
D
(
mc
i
)=
D
for
i
=1
,...,l
, and let
m
i
=
β
(
w
(
mc
i
)
,cs
(
mc
i
))
. Then the
motivation value
μ
(
D
)
for a desire
D
is defined as
μ
(
D
)=
f
(
m
1
,f
(
m
2
,...,f
(
m
l−
1
,m
l
)
...
))
for
l>
1
and
μ
(
D
)=
m
1
otherwise.
