Struct ConstraintSolver 

pub struct ConstraintSolver {
    pub variables: Vec<CspVariable>,
    pub constraints: Vec<Constraint>,
    pub config: SolverConfig,
}

A full-featured CSP solver combining AC-3 domain pruning with backtracking search and configurable variable/value ordering heuristics.

Fields

variables: Vec<CspVariable>

All registered variables (in insertion order).

constraints: Vec<Constraint>

All registered constraints.

config: SolverConfig

Solver configuration.

Implementations

impl ConstraintSolver

pub fn new(config: SolverConfig) -> Self

Create a new, empty solver with the given configuration.

pub fn add_variable(&mut self, name: String, domain: Vec<i64>) -> CspVarId

Register a new variable with the given name and initial domain values.

Returns the CspVarId that uniquely identifies this variable.

pub fn add_constraint(&mut self, c: Constraint)

Append a constraint to the problem.

pub fn stats(&self) -> CspStats

Return a summary of the current CSP instance.

pub fn is_consistent( &self, assignment: &Assignment, var: CspVarId, value: i64, ) -> bool

Return true when all constraints that are fully grounded by assignment (plus the tentative assignment of value to var) are satisfied.

pub fn check_constraint( &self, c: &Constraint, assignment: &Assignment, var: CspVarId, value: i64, ) -> bool

Check a single constraint under a partial assignment with var = value.

Returns true when:

• the constraint cannot yet be evaluated (some participant is unassigned), OR
• the constraint is satisfied by the current (partial+tentative) assignment.

pub fn ac3(&self, domains: &mut [Domain]) -> bool

Run the AC-3 arc-consistency algorithm on the given domains (which are cloned from the initial variable domains at the start of solve).

Returns false when arc consistency reveals that the problem has no solution (some domain became empty).

pub fn revise(&self, domains: &mut [Domain], xi: CspVarId, xj: CspVarId) -> bool

Revise xi’s domain: remove any value that has no support in xj.

Returns true when at least one value was removed from xi’s domain.

pub fn select_unassigned_variable( &self, assignment: &Assignment, domains: &[Domain], ) -> Option<CspVarId>

Select the next unassigned variable to expand.

When use_mrv is enabled, picks the variable with the fewest remaining domain values (ties broken by variable id). Otherwise returns the first unassigned variable in insertion order.

pub fn order_domain_values( &self, var: CspVarId, _assignment: &Assignment, domains: &[Domain], ) -> Vec<i64>

Return the values from var’s current domain in the order we should try them during backtracking.

When use_lcv is enabled this would apply the Least Constraining Value heuristic; currently we return values in sorted (ascending) order as a lightweight deterministic approximation.

pub fn backtrack( &self, assignment: &mut Assignment, domains: &mut Vec<Domain>, result: &mut SolverResult, ) -> bool

Recursive backtracking search.

Returns true when the search should stop (either max_solutions reached or max_backtracks exceeded).

pub fn solve(&mut self) -> SolverResult

Solve the CSP and return a SolverResult with all found solutions and search statistics.