Game Development Reference
In-Depth Information
The reason that we, as intelligent people, find this behavior absurd is that we
recognize the futility of the situation. In fact, the popular (and generally very accu-
rate) layman's definition of
insanity
is “doing the same thing over and over and ex-
pecting a different result.� Given that one of the early structural flaws about which
beginning programmers are cautioned is the “infinite loop,� most of us are very
cognizant of the fact that computers are blissfully unaware of the existence of, much
less the absurdity of, futility. Unless instructed otherwise or until a different com-
plication (such as your processor chip melting) brings it to a halt, a computer will
contentedly repeat mundane tasks
ad infinitum
.
One of our tasks, then, is to make sure our AI agents
understand
the concept of
futility. We can do this through pattern recognition or a variety of other esoteric
techniques. However, it is usually better to avoid the situation altogether by provid-
ing the agent with more information and a better decision model.
B
UILDING
D
ECISION
M
OMENTUM
We often refer to the antithesis of strobing as
decision momentum
. The use of the
word
momentum
is not entirely figurative. It actually is appropriate in a metaphor-
ical sense. We use the word in the same sense as we use it in physics. In the world
of Sir Isaac Newton, when something is barreling along in one direction, it is going
to take something pretty big to get it to change direction, much less reverse its di-
rection entirely.
Remember that in Newtonian physics, mass is only part of the equation. The
momentum of an object is a product of its mass
and
its velocity. When two objects
of equal mass meet in a collision, the one that is moving will “win� against the one
that it is at rest. For example, imagine a moving billiard ball colliding with a station-
ary one. Despite their equal mass, the velocity of the system continues in the direc-
tion that the original, moving ball was going prior to the collision. However, if the
moving billiard ball were to collide with something more massive such as a bowl-
ing ball (to extend things to easily illustrative lengths), the larger (albeit stationary)
mass is going to be rather convincing that the hapless billiard ball needs to head
back in the other direction. To sum up, the bowling ball had enough weight to re-
verse the direction of the entire system.
We can apply this to our choices as well. Once we have set about doing some-
thing, we have added velocity to our decision. (This is not a stretch, really. Don't we
say that we have “put a plan in motion?�) We can elect to continue performing that
action—
unless and until
something “big enough� comes along to change our direc-
tion. Because information is the impetus that puts our decisions into motion in the
first place, it is reasonable to assume that information is the force that would act
upon our decisions to get them to change their course.
