ABSTRACT
This chapter describes the principles and architecture of the Behavior Decision System (BDS). At its most basic level, the BDS is a framework which allows gameplay designers to impart knowledge to AI characters. Behavior snippets are the fundamental unit on which the BDS operates. When an AI character equips the weapon, these snippets will be registered to the character through the BDS. The BDS is based upon the assumption that it is possible to quantify the utility of every action. In the broadest possible terms, the purpose of an evaluation tree is simply to produce a score value for its associated behavior snippet. These scores can then be used as a basis for comparing two behavior snippets against each other in order to rank their relative utility. The chosen target of ability can greatly affect the outcome of executing that ability. Target selection is a necessary part of the evaluation step.
