The Gameplay Value of Learning AI in RTS Games
By: Craig Chanslor
Supervisor: Jeff Wofford
Masters of Interactive Technology degree conferred March 1, 2008
Thesis / Project completed: March 1, 2008

In recent years, technological advances in hardware have provided more raw-processing power available for gaming applications in the home. With the introduction of the next generation of consoles, graphics hardware and shader models, more graphics processing is being done on the graphics hardware, freeing processing power on the CPU for other uses. The newest generation of hardware has also introduced multi core processors into the home, allowing for threading of applications to make full use of the processing power available.

Because of this increase in processing power available, more processing power is becoming available for use in improved artificial intelligence in games. This additional processing power is allowing game AI programmers to implement more complicated AI algorithms today than in years past.  These more powerful algorithms can be advantageous in created more immersive and interesting gameplay, as the actions AI agents in the game world become more believable. These algorithms can also allow for AI agents that can learn as a game progressive, providing for either increasing more challenging enemies, or increasingly more helpful allies, to the player.  However, as more AI techniques become possible and are used in games, it gives rise to the question of what AI techniques create the best gameplay experience.

Download entire thesis (.pdf)