The moat-and-castles problem described in the paper.  Digging
increases the depth of the moat up to depth of 4.  Building a castle
has success probability that depends on the depth of the moat.
Motivated by the moat-and-castles domain of Littman and Majercik.

By the way, everything up to the first parenthesis is a comment...

(operator dig-moat
 (params (index <i>))
 (preconds (no-moat <i>))
 (effects (1.0 (del (no-moat <i>)) (add (depth1-moat <i>)))))

(operator dig-moat
 (params (index <i>))
 (preconds (depth1-moat <i>))
 (effects (1.0 (del (depth1-moat <i>)) (add (depth2-moat <i>)))))

(operator dig-moat
 (params (index <i>))
 (preconds (depth2-moat <i>))
 (effects (1.0 (del (depth2-moat <i>)) (add (depth3-moat <i>)))))

(operator dig-moat
 (params (index <i>))
 (preconds (depth3-moat <i>))
 (effects (1.0 (del (depth3-moat <i>)) (add (depth4-moat <i>)))))

(operator build-castle
 (params (index <i>))
 (preconds (no-moat <i>))
 (effects (0.25 (add (castle <i>))) (0.75)))

(operator build-castle
 (params (index <i>))
 (preconds (depth1-moat <i>))
 (effects (0.45 (add (castle <i>))) (0.55)))

(operator build-castle
 (params (index <i>))
 (preconds (depth2-moat <i>))
 (effects (0.60 (add (castle <i>))) (0.40)))

(operator build-castle
 (params (index <i>))
 (preconds (depth3-moat <i>))
 (effects (0.70 (add (castle <i>))) (0.30)))

(operator build-castle
 (params (index <i>))
 (preconds (depth4-moat <i>))
 (effects (0.75 (add (castle <i>))) (0.25)))

