Authors: Sanish Sampath, Duncan Coulter, and Elizabeth Ehlers
In this paper, the proposition and the results of the development of a novel model and associated prototype are conveyed in the field of Artificial Intelligence. The model proposed is that of agent model for a non-player character (NPC) that attempts to make the NPC more human-like and less predictable to players of the game environment. To accomplish this the model introduces a basic learning methodology with influence from that of behavioral psychology that targets the rate of learning and the decision-making process. The structure of the NPC agent model is fashioned based on the concept of symbiotic game agent. The environment proposed to feature the model is a quiz game which is a form of serious game. The results produced from this can be considered satisfactory and that the model meets its objectives. More tests in different environment of game types would aid in the further development of this form of psychological modelling. The areas other than that of game development that the model can be utilized in is that of acting as a monitoring agent in the field of Information Security.
Keywords: Artificial Intelligence; serious games; psychology; symbiotic game agents; learning