BDI player modelling for interactive narratives

Rivera Villicana, J 2018, BDI player modelling for interactive narratives, Doctor of Philosophy (PhD), Science, RMIT University.


Document type: Thesis
Collection: Theses

Attached Files
Name Description MIMEType Size
Rivera_Villicana.pdf Thesis application/pdf 3.25MB
Title BDI player modelling for interactive narratives
Author(s) Rivera Villicana, J
Year 2018
Abstract In this thesis, we present an approach to model player behaviour in Interactive Narratives (INs). INs are a form of storytelling where the listener can affect the direction of the story by making choices at different stages. As the use of INs in computer games becomes popular to provide a rich experience, the interest in predicting player behaviour and preferences in such an environment rises. This knowledge could aid in retaining a customer base, producing realistic character behaviour, or better writing and quality assurance processes. This work investigates the use of the BDI framework and apprenticeship learning to tackle the limitations involved in emulating human behaviour in INs. Since INs are partially observable for the player, their behaviour is often sub-optimal. In addition, the difference in styles and preferences results in different behaviour among players. Finally, besides modelling the knowledge regarding the state of the environment and preferences of the player, there is a need to model and simulate “common sense” human decision making for scenarios where the game state and player’s preferences are irrelevant. Following an iterative design process, we created and enhanced BDI and apprenticeship learning agents with the aim to mimic the behaviour of different players. Modelling human players as artificial agents required the acquisition and representation of knowledge by studying their behaviour with qualitative and quantitative methods between iterations. This research contributes to the area of artificial intelligence applied to games with INs. The specific contributions presented are methodologies to build BDI models to emulate generic and/or specific human behaviour; a player profile with elements that define the difference in behaviour among players; an analysis of players’ behaviours, goals and strategies; and an apprenticeship learning model that encodes the behaviour of players.
Degree Doctor of Philosophy (PhD)
Institution RMIT University
School, Department or Centre Science
Subjects Virtual Reality and Related Simulation
Adaptive Agents and Intelligent Robotics
Keyword(s) player modelling
interactive narratives
BDI
apprenticeship learning
belief-desire-Intention
user modelling
Versions
Version Filter Type
Access Statistics: 60 Abstract Views, 90 File Downloads  -  Detailed Statistics
Created: Wed, 31 Jul 2019, 14:46:45 EST by Keely Chapman
© 2014 RMIT Research Repository • Powered by Fez SoftwareContact us