Computational techniques for reasoning about and shaping player experiences in interactive narratives

Interactive narrative experiences are marked by two important characteristics: (1) a large space of player interactions, some subset of which are specified as aesthetic goals for the system to realize and (2) the affordance for the player to express self-agency and interact in a meaningful way with the system. As a function of those characteristics, players are (often unknowing) participants in the creation of the narrative experience. They cannot be assumed to be cooperative, nor can they be assumed to be adversarial. Thus, we must provide paradigms to designers that enable them to work with the players to co-create the experiences without the explicit transfer of the system's goals (as specified by the author) to players and without the need for the system to have a goal-oriented model of a player's behaviors or decisions. This dissertation formalizes compact representations and efficient algorithms that enable computer systems to represent, reason about, and shape player experiences in interactive narrative settings.Early work in the field of interactive narrative relied heavily on so-called "script-and-trigger" systems, requiring sizable engineering efforts from designers to provide concrete instructions for when and how systems can modify or adapt an environment in service of providing a narrative experience to a player. While there have been advances in techniques for representing and reasoning about narrative experiences at an abstract level that automate (to a degree) the trigger side of script-and-trigger systems, few techniques have succeeded in providing designers with a paradigm that reduces their need for scripting system adaptations or reconfigurations---one of the contributions of this dissertation.We first describe a decomposition of the design process for interactive narrative that induces three distinct technical problems: goal selection, action/plan selection/generation, and action/plan refinement. This decomposition allows technical machinery to be developed for reasoning about the complete implementation of an interactive narrative. In turn, we describe representational and algorithmic solutions to each of these problems: a Markov Decision Process-based formalism for goal selection, a schema-based planning architecture based on theories of influence and persuasion from social psychology for action/plan selection/generation, and a natural language-based template system for action/plan refinement. To evaluate these techniques, we conduct experiments both in simulation and in an interactive storytelling system with human subjects.In using these techniques, we realize the following three goals: (1) provide efficient algorithmic support for authoring interactive narratives for entertainment, education, or training (2) design a paradigm for AI systems to reason about and act to shape player experiences based on author-specified aesthetic goals for interactive narratives and (3) accomplish (1) and (2) with the player feeling more engaged and without the player perceiving a decrease in self-agency.