Abstract
The digital gaming industry is a competitive industry. The success of a game is often dependent on the number of copies sold or the number of times the game is replayed. Player experience is another indicator of game success. Creating a good player experience is achievable through creating dynamic and rational non-player characters (NPCs). The AdaptiveSGA model has been shown to create dynamic and rational non-player characters. In addition, the AdaptiveSGA model extends and simplifies the Symbiotic Game Agent (SGA) model. This dissertation aims to determine if the competitiveness of the AdaptiveSGA model can be improved by knowledge sharing concepts and superposition principles specified in Superposition Collaborative Multi-Agent Systems Architecture (SPLINTER). The dissertation attempts to answer this question by introducing the Knowledge Sharing Symbiotic Game (KSSG) model. The KSSG model extends the AdaptiveSGA model by introducing knowledge sharing concepts and SPLINTER superposition principles. The KSSG model leverages knowledge sharing concepts by the transfer of knowledge from past generations. Furthermore, the KSSG model implements SPLINTER superposition principles to swap agents during runtime to adaptively change NPC competitiveness during a game. SPLINTER’s superposition principles were designed by considering quantum superposition. The KSSG model consists of two principal components: the KSSG architecture and the KSSG agent. The KSSG architecture is comprised of KSSG agents. The KSSG model was tested in a prototype cricket game called KSSG-Action Cricket and was shown to improve the competitiveness of the AdaptiveSGA model by the introduction of knowledge sharing concepts and SPLINTER superposition principles.