Granularity in multi-method planning

Authors:
Soowon Lee;Paul S. Rosenbloom
Affiliations:
Information Sciences Institute and Computer Science Department, University of Southern California, Marina del Rey, CA;Information Sciences Institute and Computer Science Department, University of Southern California, Marina del Rey, CA
Venue:
AAAI'93 Proceedings of the eleventh national conference on Artificial intelligence
Year:
1993

Citing 5
Cited 0

Depth-first iterative-deepening: an optimal admissible tree search

Artificial Intelligence
Planning for conjunctive goals

Artificial Intelligence
SOAR: an architecture for general intelligence

Artificial Intelligence
Partial-order planning: evaluating possible efficiency gains

Artificial Intelligence
Learning effective search control knowledge: an explanation-based approach

Learning effective search control knowledge: an explanation-based approach

Quantified Score

Hi-index	0.00

Visualization

Abstract

Multi-method planning is an approach to using a set of different planning methods to simultaneously achieve planner completeness, planning time efficiency, and plan length reduction. Although it has been shown that coordinating a set of methods in a coarse-grained, problem-by-problem manner has the potential for approaching this ideal, such an approach can waste a significant amount of time in trying methods that ultimately prove inadequate. This paper investigates an approach to reducing this wasted effort by refining the granularity at which methods are switched. The experimental results show that the fine-grained approach can improve the planning time significantly compared with coarse-grained and single-method approaches.