Approaches to Improving Believability and Effectiveness of Agents in Video Games

One constant issue in development of video games is the tremendous effect that game agents have on the quality of the player's experience. In many modern games, a single player or even multiple players are given an opportunity to have bots as opponents or as team-mates. In addition, some games involve players interacting with friendly non-player characters (NPCs) who are not companions. Interactions with such characters vary based on the game genre and the implementation of the individual game. However, one common goal among most game genres and character types, is to have effective artificial intelligence (AI) that keeps the player involved and enthusiastic. This means that in games involving combat or sports, team-mates must be responsive and must not hinder accomplishment of a goal and opponents must be non-predictable. In more cinematic role-playing games (RPGs), both friendly and enemy characters require dimension and a realistic forum in which to interact with the player for the experience to be satisfying. Work has been done in researching the believability of characters in video games, along with several recent approaches to implementing team-based AI. There are also many ways to model characters, in addition to algorithms which allow game agents to be less predictable and thus, more enjoyable to play with. Studies conducted on the quality and methodology of game agent implementation continue to aid the video game industry in enriching the experience for all types of players.