Authors: Mark Fiala
AI (Artificial Intelligence) behaviors of automated elements of a video game are an important part of the user’s experience. Often the “intelligence” of the automated elements, usually enemies, is a major
feature affecting user acceptance. Multi-agent systems, such as action-strategy games, have a special need for AI behaviors that function competently. Typically AI behavior design has been ad hoc with unexpected outcomes and delays in the product design schedule. Usually the problems are due to too many feedback paths built one on top of another to try achieve a desired result. This paper presents a simple paradigm and design methodology to achieve predictable and desired behaviors in a reasonable time frame with even a beginner or intermediate programmer.
Keywords: AI behaviors, multi-agent planning, action-strategy games.