Struct MetricRegularityChecker 

pub struct MetricRegularityChecker {
    pub radius: f64,
    pub num_tests: usize,
    pub tol: f64,
}

Checks metric regularity properties of a mapping at a point.

Fields

radius: f64

Perturbation radius for numerical testing.

num_tests: usize

Number of test perturbations.

tol: f64

Tolerance for regularity bound checks.

Implementations

impl MetricRegularityChecker

pub fn new(radius: f64, num_tests: usize, tol: f64) -> Self

Create a new checker.

pub fn estimate_modulus( &self, f: impl Fn(&[f64]) -> Vec<f64>, x0: &[f64], y0: &[f64], domain_samples: &[Vec<f64>], ) -> f64

Estimate the metric regularity modulus κ of F at (x₀, y₀): κ ≈ sup_{(x,y) near (x₀,y₀)} d(x, F⁻¹(y)) / d(y, F(x)).

Here F is represented as f: &[f64] -> Vec<f64> and F⁻¹(y) is approximated by the preimage computed via nearest-point search.

pub fn check_constraint_qualification( constraints: &[impl Fn(&[f64]) -> f64], x: &[f64], h: f64, ) -> bool

Check whether a constraint qualification (MFCQ-like) holds by verifying that the constraint gradients at x span a “rich” direction set.

pub fn check_quasiconvex( f: impl Fn(&[f64]) -> f64, x: &[f64], y: &[f64], num_samples: usize, ) -> bool

Check whether a set of points satisfies a quasiconvexity condition: f on the line segment [x,y] is ≤ max(f(x), f(y)) for sample points.