Show simple item record

dc.contributor.authorVattam, Swaroop Srinivasa Murthy
dc.description.abstractThe central theme of this research is simple: people are able to cope with complexities of the real world because of their ability to improvise, which makes them more flexible and adaptable. Most artificial intelligence systems are rigid, do not scaleup to challenges of the real world, and do not exhibit improvisational behavior. If we acknowledge the benefits of improvisation methods for people, then we can also acknowledge that similar benefits apply to artificial intelligence systems. Therefore, if we can build systems that can improvise we are better-off. Based on this simple theme, we have proposed a novel architecture that combines a traditional planner with an analogy-based improvisation component in order to craft planning systems that respond intelligently in complex and little-known environments. CREAP (CREative Action Planner) is an intelligent agent embodying our architecture. Situated in a simulated environment of its own, CREAP is designed to plan and act in ways that achieve certain prespecified goals. Experiments conducted on CREAP show that our architecture is capable of producing improvisational behavior, thus lending flexibility to our agent in responding to less familiar situations.
dc.subjectArtificial Intelligence
dc.subjectAnalogical reasoning
dc.subjectTestbed systems
dc.titleAn architecture for analogy-based plan improvisation
dc.description.departmentArtificial Intelligence
dc.description.majorArtificial Intelligence
dc.description.advisorDonald E. Nute
dc.description.committeeDonald E. Nute
dc.description.committeeWalter D. Potter
dc.description.committeeElizateth Preston

Files in this item


There are no files associated with this item.

This item appears in the following Collection(s)

Show simple item record