Narrative Planning: Compilations to Classical Planning
This work addresses the challenge of efficient story generation for AI applications, but it is incremental as it builds on existing planning methods.
The paper tackles the problem of story generation by showing that the intentional behavior condition in narrative planning can be compiled away to create a classical planning problem, which is solved more efficiently using an off-the-shelf planner compared to a specialized one.
A model of story generation recently proposed by Riedl and Young casts it as planning, with the additional condition that story characters behave intentionally. This means that characters have perceivable motivation for the actions they take. I show that this condition can be compiled away (in more ways than one) to produce a classical planning problem that can be solved by an off-the-shelf classical planner, more efficiently than by Riedl and Youngs specialised planner.