Information Technology Reference
In-Depth Information
In the second experiment, we have also established a dominant drive in each prob-
lem and the agent has three objects of each type. In this case, the targets are again
getting a value below a threshold. But, we added another goal: to achieve a valence
value above a threshold, so that we can consider that the agent is in a good state af-
ter the execution of the plan. We have established different standard durations for ac-
tions: instantaneous (as
take
and
leave
objects), short duration (as
drink
,
go
and
shower
), medium duration (as
eat
,
bath
,
play
,
watch
,
read
,
listen
and
buy
)
and long duration (
sleep
). The metric was minimizing the total duration of the plan.
The experiments were performed with two different personalities according to the con-
scientiousness value: careless and meticulous, because conscientiousness weights the
standard duration of the actions.
Fig. 5.
Quality of the plans for the stable agent
Fig. 6.
Quality of the plans for the neurotic agent
