Traversing Trees

• The control part basically implements a search procedure
• Parts of the steps in the search also add / remove / change constraints
• The search tree is formalized as a transition system between states
• Each node (state) in the search tree contains:
  • The goals which remain to be solved
  • The constraints the constraint solver can handle
  • The constraints the constraint solver cannot handle (yet)
• States are represented as a 3-tuple $\langle{A, C, S}\rangle$
• An initial state is $\langle{G, \emptyset, \emptyset}\rangle$
• Final state: $E \longrightarrow E$
• A successful derivation: final state $\langle{\emptyset, C, S}\rangle$