Enum TerminationReason 

#[non_exhaustive]
pub enum TerminationReason {
    MaxIter,
    MaxCostEvals,
    MaxGradientEvals,
    GradientTolerance,
    RelativeGradientTolerance,
    ProjectedGradientTolerance,
    ParamTolerance,
    RelativeParamTolerance,
    CostTolerance,
    RelativeCostTolerance,
    TargetCost,
    NoImprovement,
    SimplexTolerance,
    CmaEsTolerance,
    RhoTolerance,
    MeshTolerance,
    MaxTime,
    SolverConverged,
    SolverFailed,
}

Why the executor stopped. Returned on OptimizationResult::reason and the various step/run hooks.

Variants (Non-exhaustive)

Non-exhaustive enums could have additional variants added in future. Therefore, when matching against variants of non-exhaustive enums, an extra wildcard arm must be added to account for any future variants.

MaxIter

state.iter() >= max_iter.

MaxCostEvals

Cost-evaluation budget exhausted.

MaxGradientEvals

Gradient-evaluation budget exhausted.

GradientTolerance

‖∇f(x)‖ ≤ tol.

RelativeGradientTolerance

‖∇f(x_k)‖ ≤ tol · ‖∇f(x_0)‖: gradient norm relative to the initial gradient (scale-invariant first-order stationarity).

ProjectedGradientTolerance

‖x − π_C(x − ∇f(x))‖_∞ ≤ tol: projected-gradient stationarity for box-constrained problems. Collapses to the unconstrained gradient norm when no constraint is active.

ParamTolerance

‖x_k − x_{k−1}‖ ≤ tol.

RelativeParamTolerance

‖x_k − x_{k−1}‖ ≤ tol · ‖x_k‖: scale-invariant step test (MINPACK xtol).

CostTolerance

|f_k − f_{k−1}| ≤ tol.

RelativeCostTolerance

|f_k − f_{k−1}| ≤ tol · |f_{k−1}|: scale-invariant cost reduction test (MINPACK ftol).

TargetCost

f(x_k) ≤ target: user-supplied target cost reached (NLopt’s stopval/SciPy’s f_min).

NoImprovement

Best-so-far cost has not improved by more than tol in patience consecutive iterations: the early-stopping pattern.

SimplexTolerance

Simplex collapsed below the configured tolerance.

CmaEsTolerance

CMA-ES search distribution collapsed below TolX: σ · maxᵢ dᵢ < tol_x (Hansen 2016 Appendix B.3).

RhoTolerance

NEWUOA trust-region radius reached the configured floor: ρ ≤ rho_end.

MeshTolerance

MADS poll size reached the configured floor: Δᵖ ≤ poll_size_min.

MaxTime

Wall-clock time limit reached.

SolverConverged

Solver determined it has converged (e.g. fixed point reached).

SolverFailed

Solver cannot make further progress (e.g. line search failure).

Implementations

impl TerminationReason

pub fn is_failure(&self) -> bool

Whether this reason represents an unrecoverable failure that an outer solver should bubble (rather than consume and continue).

Currently only SolverFailed qualifies: MaxIter, the *Tolerance reasons, and SolverConverged are all “clean stops” that an outer solver running an inner per outer iter should treat as “result is fine, move on”. See CONTRIBUTING.md “Solver composition” for the failure-routing contract.

Trait Implementations

impl Clone for TerminationReason

fn clone(&self) -> TerminationReason

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Copy for TerminationReason

impl Debug for TerminationReason

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Eq for TerminationReason

impl PartialEq for TerminationReason

fn eq(&self, other: &TerminationReason) -> bool

Tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl StructuralPartialEq for TerminationReason