Constraint Programming with CP Flashcards

In Picat CP, what does `fd_dom(Var, Dom)` return?

Answer: The domain of Var as a list of values or ranges

`fd_dom(Var, Dom)` unifies Dom with a representation of all remaining values in Var's finite domain.

Which constraint ensures that the values of variables in a list form a permutation of 1..N in Picat CP?

Answer: all_different(List) with domain 1..N

Combining `all_different(List)` with each variable's domain set to `1..N` exactly models a permutation constraint.

What happens if you post two contradictory constraints, such as `X #= 3` and `X #= 5`, on the same variable in Picat CP?

Answer: The predicate fails, triggering backtracking

Constraint propagation detects the empty domain immediately and causes the current goal to fail, which triggers backtracking in the search.

In Picat's CP module, the `circuit/1` global constraint is used to model which type of problem?

Answer: Hamiltonian cycle over a set of successor variables

`circuit(Succ)` constrains the array Succ so that the successor relation forms a single Hamiltonian cycle visiting every node exactly once.

Which predicate tests whether a CP variable is already instantiated (ground) in Picat?

Answer: ground(Var)

`ground(Var)` succeeds if Var is fully instantiated and contains no unbound logical variables.

In Picat CP, what does the `max_regret` variable-selection heuristic do?

Answer: Chooses the variable whose two best values differ most, maximizing the cost of wrong choice

`max_regret` selects the variable where the difference between the best and second-best domain values is greatest, focusing on high-stakes decisions.

What is the role of `scalar_product/4` in Picat's CP module?

Answer: Models a linear constraint between a coefficient vector and variable vector

`scalar_product(Coeffs, Vars, Rel, Bound)` posts the linear constraint Coeffs · Vars Rel Bound using the specified relational operator.