Quadratic-form bounds on subspaces #
Scalar-generic lower and upper bounds for the real part of the quadratic form of a bounded operator, restricted to a subspace. These predicates are useful well beyond Davis--Kahan perturbation theory.
Provenance #
- Original repository: Davis--Kahan/DKPS formalization (Kitware, Inc.).
- Original module:
ForMathlib/Analysis/InnerProductSpace/QuadraticFormBounds.leanin the DKPS formalization at commitdf036cd. - Authors: Jon Crall and GPT 5.6 High; Copyright (c) 2026 Kitware, Inc.; released under Apache 2.0.
Form-bound predicates extend the ContinuousLinearMap API and support
dot notation on bounded operators.
Lower quadratic-form bound on a subspace.
Instances For
Upper quadratic-form bound on a subspace.
Instances For
The characteristic inequality defining a lower form bound.
The characteristic inequality defining an upper form bound.
Basic theory #
The two ways a form bound weakens -- in the constant and in the subspace -- and the
identification of the degenerate case with Mathlib's IsPositive. A consumer holding a
bound on U at constant c and needing one on a subspace of U, or at a worse constant,
should not have to reprove it from the definition.
A lower form bound weakens as the constant decreases.
A lower form bound restricts to a smaller subspace.
An upper form bound weakens as the constant increases.
An upper form bound restricts to a smaller subspace.
Positivity implies the zero lower form bound on the whole space.
The converse: symmetry plus the zero lower bound on ⊤ is positivity. Together with
IsPositive.lowerFormBoundOn_top this pins LowerFormBoundOn _ ⊤ 0 as a generalization of
Mathlib's predicate rather than a competitor to it.
The real quadratic form of a real scalar operator is the scalar times the squared norm.