This work extends Koehler and Hoffmann's definition of reasonable orders between top level goals to the more general case of landmarks and shows how landmarks can be found, how their reasonable orders can be approximated, and how this information can be used to decompose a given planning task into several smaller sub-tasks.
An approach to planning with trajectory constraints is developed that decomposes the problem into a set of smaller subproblems using the temporal orderings described by the constraints and then solves them incrementally.
This work defines ordering constraints not only over the top level goals, but also over the sub-goals that will arise during planning, and demonstrates that the approach can yield significant performance improvements in both heuristic forward search and Graphplan-style planning.
It is shown how the use of state constraints can provide a unified perspective on important problems faced in IS and the development of an approach to narrative generation that exploits such constraints are developed.
An approach for learning planning domain models directly from natural language (NL) descriptions of activity sequences, which starts from NL descriptions of actions and uses NL analysis to construct structured representations, from which to construct formal representations of the action sequences.
A new approach to the definition of virtual characters aimed at achieving a balance between character autonomy and global plot structure is reported and the notion of a characters' Point of View is introduced and shown how it enables a story to be described from the perspective of a number of different characters.
It is argued that the structure of social relationships between characters can be used as a powerful mechanism to determine a narrative, putting less emphasis on the details of plot structure and closer to how modern dramas are shaped in specific genres, where situations and relationships are determinant.
This work replaces single intentional planners with multiple agents representing the characters of a narrative, which can reason about the relevance of narrative actions given their individual intents, using a state-based forward search procedure that results in a significantly smaller search space.
Work in progress to extend some features of domain analysis using landmarks, including the extraction of resource abstracted landmarks; and the identification of landmark repetition along with a count of the minimum number of times a landmark will need to be repeated.