Documentation

LeanCert.Tactic.IntervalAuto.Bernstein

Bernstein polynomial bounds #

bernstein_bound proves ∀ x ∈ I, f x ⋈ c for a univariate rational polynomial f by the recursive Bernstein certificate LeanCert.Validity.Bernstein.checkPoly*Bound. Untrusted search finds the least bisection depth at which the certificate succeeds; the retained proof closes that single Boolean certificate through the configured verification route and lifts it with the matching golden theorem.

The strategy is typed and transactional: every non-success restores the complete caller state, and a goal that is not a polynomial reports notPolynomial so the semantic router can skip it without spending budget.

Runtime facts retained by a successful Bernstein certificate.

Instances For

    Expected non-successes of the Bernstein strategy.

    Instances For
      Equations
      • One or more equations did not get rendered due to their size.
      Instances For
        Equations
        • One or more equations did not get rendered due to their size.
        Instances For

          Untrusted: recognise the reified AST as a QPoly.

          Equations
          • One or more equations did not get rendered due to their size.
          Instances For

            Untrusted: is the reified AST a univariate rational polynomial? Used by the router to decide applicability before spending budget.

            Equations
            • One or more equations did not get rendered due to their size.
            Instances For

              Typed and exception-total Bernstein entry point. Every non-success restores the complete caller tactic state. Success preserves pending sibling goals.

              Equations
              • One or more equations did not get rendered due to their size.
              Instances For

                bernstein_bound [depth] proves a univariate polynomial interval bound by Bernstein coefficients, bisecting at most depth times (default 8).

                Equations
                • One or more equations did not get rendered due to their size.
                Instances For