Artificial Mind for Virtual Characters - Advancing Artificial Intelligence Through Biological Process Applications - page 193

be included in the Cognition Agent activated
by the Motivation Agent. It corresponds to the
Frustration Attitude of the character toward that
particular motivation. This procedure includes
two routines: the activation of a special Execution
Agent (the Discharge Agent ), and the update of
the Motivation Intensity value i , by returning it
to the T max value. A Discharge Execution Agent
may execute, for instance, a meaningless behavior,
like walking in circles in the room expressing
impatience. So, the Discharge Procedure may be
as simple as this:

Motivation deactivation d occurs when specific
actions take place, either satisfying the motivation
or just discharging frustration. The Cognition
Agent that was activated by the Motivation Agent
sends the corresponding value.
Intrinsic autonomous rate growth m represents
the self-generated component of a motivation.
It codes the relative importance that the cor-
respondent Intention has to the character in
question. Basic natural needs (hunger, thirsty,
etc.) probably have the same m for all characters
of a single environment, except when there is a
pathologic state associated with anyone of them
(like anorexia, for instance). Complex psychologi-
cal drives (fear, aggressivity, etc.) tend to vary
between different characters, as they reflect their
particular traits.
When a Motivation Agent wins the competition
for priority, it triggers the activation of a chain of
agents that eventually produce the Motivational
deactivation value d , which represents the amount
of discharge obtained with the performed action.
It may not be enough to take the intensity to the
comfort zone. For instance, if a character is very
hungry, but the only food obtainable in the en-
vironment is, say, an apple, it probably will not
be sufficient to completely satisfy the hunger,
although it may decrease it a little. In this case,
d value is not high enough, and the Hunger agent
will still compete for priority in the next cycle.
A special case is when a motivation can only
be activated by an external stimulus. For instance,
usually people feel fear only when confronted
with something scary, and then the prior goal
becomes Safety . So, the normal state of the
motivation agent Safety is i = 0 and m = 0, when
nothing threatening is being perceived, and there
is no external reason to feel threatened. But, by
definition, Motivation Agents don't have access
to the outer environment. So, there has to be a
special Cognition Agent that sends an alert sign
to the Safety agent. This special case of Cogni-
tion Agent is called Alert Agent, and it is detailed
in next section. For now, it is important to note

IF i > T max
THEN
{Activate Discharge Execution Agent;
i = T max }

Updating i by returning it to the T max value
keeps the motivation in the competition for priority.
So, it can be satisfied later, when the opportunity
arrives. It is possible to make the character give
up trying, if the procedure returns the Intensity
value to the comfort zone after a certain limit of
time, for example. The context may have to be
analyzed to decide which discharges are possible,
and which are not. Anyway, it is important to deal
with the frustration, and this can be done with
small variations of this simple procedure.
Since every Motivation Agent begins with i =
0 (which means that no motivation is active), there
has to be an internal procedure that changes this
value to reflect the variation of each particular
motivation with time. The attribute that allows
this to take place is the variable m , which repre-
sents the intrinsic autonomous rate of motivation
growth. So, let Intensity i t+1 at time t+1 be:

i t+1 = i t + m - d ,
where:
i t : intensity value at time t ;
m : intrinsic autonomous rate of motivation
growth;
d : motivational deactivation.