Decision-theoretic Refinement Planning Using Inheritance Abstraction

Abstract

ing Actions We extend Tenenberg’s (Tenenberg 1991) notion of inheritance abstraction for STRIPS operators to apply to conditional probabilistic actions. As Tenenberg explains it, "the intent of using inheritance abstraction is to formalize the notion that analogous action types can be structured together into an action class at the abstract level, characterized by the common features of all elements of the class." Thus we can plan with the abstract action and infer properties of a plan involving any of the instances of the abstract action. We abstract actions by grouping their outcomes. If one action has an outcome el and another action has an outcome e2 then we can produce an abstract action HADDAWY 267 From: AIPS 1994 Proceedings. Copyright © 1994, AAAI (www.aaai.org). All rights reserved.

Extracted Key Phrases

8 Figures and Tables

Statistics

02040'95'97'99'01'03'05'07'09'11'13'15'17
Citations per Year

139 Citations

Semantic Scholar estimates that this publication has 139 citations based on the available data.

See our FAQ for additional information.

Cite this paper

@inproceedings{Haddawy1994DecisiontheoreticRP, title={Decision-theoretic Refinement Planning Using Inheritance Abstraction}, author={Peter Haddawy and Meliani Suwandi}, booktitle={AIPS}, year={1994} }