Chapter 44. Strategic Decision Making
Key Topics
With specific capabilities in place, the focus can shift toward higher-level behaviors. This chapter combines the analysis and understanding phases of the AI development. A solid grasp of the decision-making process in the context of high-level strategies is necessary for manually reproducing human behaviors, as well as for developing a learning AI capable of adapting to deathmatch situations.
There's no need for specification of the world interfaces, because the higher levels of AI only depend on lower levels rather than senses and actuators.
The chapter covers the following topics:
Factors in tactical decisions, some of which are used as features of the problem Objectives in the game, both global and personal ones Base components of tactical behaviors that are combined The decision-making process in the context of strategies A training zone that can be used for learning and evaluation
At the end of this chapter, the necessary elements are ready for a hard-coded solution to be developed (as well as a learning AI to deal with deathmatch games).
|