Game Development Reference
In-Depth Information
We touched on this process briefly in Chapter 3 using the example of how my
children's teachers calculate grades through first combining the scores for home-
work, tests, and quizzes together into their own scores and then combining those
scores together into one aggregate score. We revisited the idea through our discus-
sion of multi-attribute utility theory in Chapter 9. Each of these examples follows
the same pattern. We start with small blocks of information that we then combine
into larger blocks. Theoretically, there is no limit to how many layers we can build
or how many pieces of information can go into any one aggregate.
Once again, we can use any of the above methods to do this. Typically, however,
weighted sums allow us to combine the component parts together in the most flex-
ible manner.
C
ONSTRUCTING A
L
AYER
The unfortunate weakness in the dinner example above is that our desire (separate
or combined) is not the
only
factor in the decision. There are other thoughts that we
must entertain. For example, we often have three factors that we want to take into
consideration.
desire
How much we desire the selection
price
The price of the selection
time
The time it takes to acquire the selection
We can address each of the three items above separately. In the previous exam-
ple, we utilized weighted sums to construct a combined utility of desire. We could
generate utilities for the price of the food and the time it would take to acquire it in
a similar fashion. We could use weighted sums, response curves, or any other
method to determine the utility of each of those factors. Regardless of
how
we cal-
culated them, we would arrive at a single utility figure for
price
and
time
. When we
have arrived at those figures, we will once again use weighted sums to combine our
preferences for the four dinner options with the other factors.
You Can't Always Eat What You Want
Just as we decided how to weight my preference for dinner against Laurie's, we must
do the same with the three factors that we are including in this layer of the decision.
Because we have a lot of game programming (or book writing) to do, we decide that
the time it takes to acquire the selection is the most important factor. Additionally,
because game programming (or writing about game artificial life [AI]) doesn't
make us rich, we decide that the price of our prospective forage is the second most
important factor. While there is
some
consideration for how much we actually
like
the food, its impact is incidental.
