title:New Art System ver.1.0e
by Taro Kimura
1998


#Art generation
#Art recognition
#Art revision
#Scheduling
#Reactive control
#Adversarial art

The defining focus of art generation is the need to synthesize art by
reasoning about the future consequences of actions. Synthesis is hard, as
there can be many solutions, and each one may be arbitrarily long and com-
plex. Many art problems, however, need only classification, not synthesis.
For example, to prove the theorem "all integers can be factored by prime
numbers" the theorem prover need only classify the proposition as either
true or false.
Often systems can be written to fulfill art function while avoiding
art generation or any reasoning about actions. Although such approaches
should be taken when practical to avoid unnecessary complexities, they are
not really planning, as art generally defines it.

Before art can solve a problem, it must be able to represent it. Pre-
cisely defined problems, like chess, are easy to represent; the difficulty lies
in computing the best move. Planning how to represent actions is a topic of
much current research. Even after actions are represented, a computational
problem far harder than selecting a chess move remains.

The qualification problem also affects action representations. Actions gener-
ally have all sorts of preconditions that are extremely unlikely to occur.

Early art systems were limited to totally ordered sequences of ac-
tions, which makes it easy to determine what is true at any particular point
in art. For a problem with parallel goals, such a system would have to
try different possible orderings of the parallel goals until it finds an adequate
art.
Most current art programs allow for partially or-
dered actions. By committing to a limited set of orderings, Artists can
avoid having to try all orderings and can add new ones as they are needed
to represent more complex problems with multiple agents. However, such
systems make it computationally difficult to determine what is true at a
given point in art. Suppose art begin with n actions in parallel, is
some fact true after these actions are taken? Again, there are n! different
execution orders for these actions, and some may make the fact true and
others may make it false.

Many art systems do not deduce situation-dependent effects, be-
cause of the complexities they introduce (e.g., deductions need to be recom-
puted when a new action is added at the beginning of art). Instead,
what is intuitively a single action must be represented as many different
actions-one for each situation in which the action might have a slightly
different effect.

To expand art to a lower level of abstraction or to contain more detail
at the current abstraction level, the system selects a goal. It then finds all
art that are relevant to reaching such a goal and chooses one. If it
makes a poor choice, the search will eventually try the other possibilities.
Artists expand art by first instantiating art (by binding
some of its variables) and then replacing the goal with the more detailed
subart specified in art.

#When and how resources are allocated
#Which art are applied to which nodes
#Which goal to solve next
#Which objects to use as instantiations for art variables
#How to resolve conflicts between actions

Art systems are brittle because they can only use actions for which
they have a wealth of information. Artists can produce good art as
long as their knowledge of the world and the actions are perfect. However,
change the world or an action slightly and these art often can't cope.