Planners are powerful tools for problem solving because they provide a complete sequence of actions to achieve a goal from a particular initial state. Classical planning research has addressed this problem in a domain-specific manner—the same algorithm generates a complete plan for any domain specification. This generality comes at a cost; domain-independent planners have difficulty with largescale planning problems. To deal with this, researchers have resorted to hand writing domain-specific planners to solve them. An interesting alternative is to use example plans to demonstrate how to solve problems in a particular domain and to use that information to automatically learn domain-specific planners that model the observed behavior. In this paper, we present the I TERANT algorithm for identifing repeated structures in observed plans and show how to convert looping plans into domain-specific template planners, or dsPlanners. Looping dsPlanners are able to apply experience acquired from the solutions to small problems to solve arbitrarily large ones. We show that automatically learned dsPlanners are able to solve largescale problems much more quickly than state-of-the-art general-purpose planners and are able to solve problems many orders of magnitude larger than generalpurpose planners can solve.
[1]
M. Fox,et al.
The 3rd International Planning Competition: Results and Analysis
,
2003,
J. Artif. Intell. Res..
[2]
Manuela M. Veloso,et al.
Analyzing Plans with Conditional Effects
,
2002,
AIPS.
[3]
Roni Khardon,et al.
Learning Action Strategies for Planning Domains
,
1999,
Artif. Intell..
[4]
Fahiem Bacchus,et al.
Planning with Resources and Concurrency: A Forward Chaining Approach
,
2001,
IJCAI.
[5]
Richard Fikes,et al.
Learning and Executing Generalized Robot Plans
,
1993,
Artif. Intell..
[6]
J. Shavlik.
Acquiring Recursive and Iterative Concepts with Explanation-Based Learning
,
1990,
Machine Learning.
[7]
Manuela M. Veloso,et al.
DISTILL: Learning Domain-Specific Planners by Example
,
2003,
ICML.
[8]
Manuela M. Veloso,et al.
Planning and Learning by Analogical Reasoning
,
1994,
Lecture Notes in Computer Science.
[9]
Jaime G. Carbonell,et al.
Learning by experimentation: the operator refinement method
,
1990
.