Crate wolfe_bfgs

Expand description

Thin BFGS-only re-export crate layered on top of opt.

Modules§

constants: Shared constants for generic backtracking and ridge-escalation policies.

AcceptedStep: A trial accepted by backtracking_line_search.
BacktrackConfig: Geometric backtracking schedule.
Bfgs: A configurable BFGS solver.
Bounds
CostStallConfig: Configuration for the gradient-independent cost-stall termination in Bfgs (enabled via Bfgs::with_cost_stall).
FiniteDiffGradient
FirstOrderSample
FusedObjective: Adapts an objective that naturally computes value and gradient together to the split ZerothOrderObjective / FirstOrderObjective interface used by the solvers.
MaxIterations
ObjectiveEvalError: A failed objective evaluation, classified by the producer.
Problem
RidgeExhausted: Exhausted result from escalate_ridge.
RidgeSchedule: Geometric ridge / Levenberg-Marquardt escalation schedule.
RidgeSuccess: Successful result from escalate_ridge.
Solution: A summary of a successful solver run.
StationarityEvidence: The quantity a stationarity claim was decided against, carried alongside the claim.
Tolerance

BfgsError: An error type for clear diagnostics.
BoundsError
ConfigError
LineSearchFailureReason
ObjectiveEvalKind: Whether a failed objective evaluation invalidates the whole run or only the point it was evaluated at.
OptimizationStatus: Outcome category for an optimizer run, distinct from the underlying Result<Solution, _> so callers can dispatch on convergence vs. budget exhaustion vs. numerical failure without pattern-matching solver-specific error variants.
Profile
StationarityKind
StationarityNorm: Which vector norm a stationarity test measured in.
StationarityScaling: Whether a stationarity threshold was absolute or rescaled by the iterate.
TerminationReason: Why a solver stopped, and the quantities that decided it.

armijo_roundoff_cushion: Round-off cushion added to an Armijo sufficient-decrease threshold.
backtracking_line_search: Evaluate geometrically contracted trial steps until one is accepted.
degenerate_trust_radius: The trust radius below which a trust region contains no point distinguishable from its centre.
escalate_ridge: Try an operation over a geometric sequence of ridge values.
optimize