posted on 2006-03-01, 00:00authored byXuemei Wang, Jaime G. Carbonell
Acquiring knowledge from experts for planning systems
is a rather difficult knowledge engineering task,
but is essential for any applications of planning systems.
This work addresses the issue of automatic acquisition
of planning operators. Operators are learned
by observing the solution traces of experts agents and
by subsequently refining knowledge in a learning-by-doing
paradigm. It is domain-independent and assumes
minimal requirements for a priori knowledge
and expert involvement in order to reduce the burden
on the knowledge engineerer and domain experts.
Planning operators are learned from these observation
sequences in an incremental fashion utilizing a
conservative specific-to-general inductive generalization
process. In order to refine the new operators
to make them correct and complete, the system uses
the new operators to solve practice problems, analyzing
and learning from the execution traces of the resulting
solutions or execution failures. We describe
techniques for planning and plan repair with incorrect
and incomplete domain knowledge, and for operator
refinement through a process which integrates
planning, execution, and plan repair. Our learning
method is implemented on top of the PRODIGY
architecture(Carbonell, Knoblock, & Minton 1990;
Carbonell et al. 1992) and is demonstrated in the
extended-strips domain(Minton 1988) and a subset of
the process planning domain(Gil 1991).