Constraint Satisfaction as Planning

Can create an artificial planning problem by introducing ``operators'' that place or move queens. Goal is to satisfy all constraints. Can use search (given operators).

Different from planning: path to goal is irrelevant.

• initial state: empty board; action: place a queen
• initial state: 8 queens down; action: move a queen

Note $n\times n$ version is solvable for all large enough n.

Is it solvable for n=1? n=2? n=3? n=4?

Note that constraints are easy to check even though it's hard to find a satisfactory configuration!