Abstract

This chapter describes the results of both generating and using abstractions for problem solving. The abstractions are generated by alpine and then used in the hierarchical version of prodigy. The chapter is divided into four sections. The first section demonstrates empirically, in the Tower of Hanoi, that abstraction can produce an exponential-to-linear reduction in search. The second section describes empirical results in several more complex domains to show that alpine can generate effective abstraction hierarchies for a variety of problem domains. The third section compares the performance of alpine’s abstractions with both hand-coded and automatically generated search control knowledge. The last section compares alpine and abstrips in the original strips domain and shows that alpine produces abstractions that have a considerable performance advantage over those generated by abstrips.

Date

September 22, 1993

Authors

Craig A Knoblock, Craig A Knoblock

Journal

Generating Abstraction Hierarchies: An Automated Approach to Reducing Search in Planning

Pages

85-106

Publisher

Springer US