Refinement (pruning) Strategies
� A plan set P is a set of partial plans {P1,P2 ... Pm}
- A refinement strategy R : P ? P� ( �P� � a subset of �P � )
- R is complete if �P� � contains all the solutions of �P �
- R is monotonic if �P� � has longer minimal candidates than �P �
- R is progressive if �P� � is a proper subset of �P �
- R is systematic if components of P� don�t share candidates
�Canned� inference procedures
-- Prune by propagating the consequences of domain theory
and meta-theory of planning onto the partial plan