Program Overview‎ > ‎

Interactive Narrative Planning using Sabre


This tutorial will introduce narrative planning in general, and present Sabre, an interactive narrative planning system. Sabre is a successor to the Glaive narrative planner, which has been used to generate and adapt stories within interactive virtual environments.

Narrative planning systems are a variant of classical AI planning systems that seek to solve a problem commonly faced by storytellers: the problem of coordinating many agents with their own goals, and who may cooperate or compete in pursuit of those goals, while reasoning about the constraints and desired outcomes of a story.

The audience will walk away with knowledge of current techniques used to solve narrative planning problems, and experience using a narrative planner.

Speaker: Stephen G. Ware

Outline of Tutorial

  1. Narrative Planning
    1. What is Narrative Planning
      1. The value of working with atomic narrative units
      2. How it is similar to classical planning
      3. How it is different from classical planning
      4. Why narrative planning is so computationally complex
      5. The problem that a narrative planner solves
    2. How Narrative Planning Works
      1. State-space search
      2. Characters goals and author goals
      3. Causal links and intentional paths
      4. Finding explanations in the search space
    3. Heuristics and Search Techniques
      1. The purpose of a planning heuristic
      2. The Fast-Forward Heuristic
      3. The Goal Graph Heuristic
      4. The Glaive Heuristic
      5. The Sabre Heuristic
      6. Basic A* and Greedy Search
      7. Enforced Hillclimbing
      8. Lookahead Search
      9. Motivation Pruning
      10. Novelty Pruning
  2. Hands-On with Sabre
    1. Domain file format
    2. Running an example project
    3. Group work time

The last half of the time will be devoted to group work time. The audience will divide up into groups, and will be provided with the Sabre Narrative Planner and a rapid prototyping tools for exploring a narrative search space as a text adventure. They will also be given a simple set of commonly used actions and a sample story. Working as a group, they will either create a new story of adapt the existing story to add new people, places, things, and actions.


Ware, S.G., Young, R.M.: Glaive: A state-space narrative planner supporting intentionality and conflict. In: Proceedings of the 10th AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment. pp. 80–86 (2014)