Abstract
Dynamic Difficulty Balancing is an essential topic in game development because it significantly contributes to the entertainment value of computer games. There are several ways of adjusting difficulty levels in computer games. The most common way is to control specific parameters, such as the health and execution speed of an enemy NPC, so that it is harder to defeat. Performing Dynamic Difficulty Balancing in this manner does not necessarily change the behaviour of NPCs but gradually decreases the ease of winning. The entertainment value of computer games can be supplemented by the inclusion of Adaptive Game AI as a mechanism for achieving Dynamic Difficulty Balancing. Adaptive Game AI enables the realisation of NPCs that display unpredictable behaviour in the process, thus making the game more enjoyable to play. This dissertation presents AdaptiveSGA, a model for implementing Dynamic Difficulty Balancing through Adaptive Game AI by modifying the Symbiotic Game Agent model. The Symbiotic Game Agent is an agent design principle based on biological symbiosis, a phenomenon that can be observed across animal and plant life on earth. The Symbiotic Game Agent model leverages the main advantage of a mutualistic symbiotic relationship: the ability of two organisms from different species to benefit each other through coexistence, thus forming a more complex organism that can achieve a goal which would not have been achievable by the individual entities on their own. Moreover, the Symbiotic Game Agent model provides flexibility and robustness because it allows for the ability to swap symbiont agents in and out of the symbiont environment without interrupting the other symbiont agents...
M.Sc. (Computer Science)