Projection-valued measures #
A projection-valued measure on the measurable sets of an arbitrary measurable
space X, acting on a complex Hilbert space H.
The finite diagonal scalar measures are provided by the DKPS donor's diag
field. Their measure-theoretic countable additivity supplies strong
countable additivity of the projections.
Idempotence, self-adjointness, positivity, and finite additivity are proved
from the structure's laws.
Main declarations #
ProjValMeasure X H: the structure — a mapprojfrom measurable sets to operators, the diagonal measuresdiag : H → Measure X, the weldinner_proj, the normalizationproj_univ : P(X) = 1, and the multiplicativity lawproj_inter.proj_idem,isSelfAdjoint_proj,isPositive_proj,proj_mul_eq_left_of_subset,proj_range_mono,proj_union,proj_compl,norm_proj_apply_le: the classical PVM laws, recovered from the structure.hasSum_proj_of_pairwise_disjoint: unconditional strong countable additivity (HasSum) on a pairwise-disjoint measurable family;tendsto_sum_proj_of_pairwise_disjointgives convergence of the natural-order partial sums.diag,diag_real_univ,diag_univ: the finite diagonal scalar measures carried by the structure, and their real/extended-real total masses.ext_of_diag,ext,ext_iff_diag,ext_iff: extensionality from either data field.
References #
- Adam Bornemann, Spectra,
Spectra/ProjValMeasure/Basic.lean, https://github.com/adambornemann-glitch/Spectra (commit8dbaaf6728d1342ae16acf79fd7eef7c59b37e63). - AIQ-Kitware, DKPS formalization,
ForTauCeti/Analysis/InnerProductSpace/ProjValMeasure/Basic.lean, https://github.com/AIQ-Kitware/aiq-dkps-formalization.
The structure #
A projection-valued measure on the measurable sets of X, acting on a
complex Hilbert space H.
The diagonal scalar measures diag ξ are supplied as finite measures,
related to the projections by inner_proj. This follows the donor's carrier
representation, in which the diagonal measures are uniquely determined by
proj. Idempotence, self-adjointness,
positivity, finite additivity, contraction, and strong countable additivity
follow from these laws.
- proj (B : Set X) : MeasurableSet B → H →L[ℂ] H
The projection assigned to each measurable set.
- diag : H → MeasureTheory.Measure X
The diagonal scalar measures, carried as data.
- diag_finite (ξ : H) : MeasureTheory.IsFiniteMeasure (self.diag ξ)
Each diagonal measure is finite (its mass is
‖ξ‖ ^ 2, bydiag_univ_toReal). - inner_proj (B : Set X) (hB : MeasurableSet B) (ξ : H) : inner ℂ ξ ((self.proj B hB) ξ) = ↑((self.diag ξ) B).toReal
The weld: diagonal matrix elements of the projections are the diagonal measures.
The whole space carries the identity.
- proj_inter (B₁ B₂ : Set X) (hB₁ : MeasurableSet B₁) (hB₂ : MeasurableSet B₂) : self.proj B₁ hB₁ * self.proj B₂ hB₂ = self.proj (B₁ ∩ B₂) ⋯
Multiplicativity: intersection of sets is composition of projections.
Instances For
The classical axioms, recovered as theorems #
Each scalar diagonal measure is finite, as witnessed by the diag_finite
field. Its total mass is computed by diag_univ.
The diagonal identity with the projection in the first inner-product slot.
The projections do not depend on the measurability witness — proof irrelevance: only the set matters.
The projection of the empty measurable set is zero.
Every measurable-set projection is idempotent.
Every measurable-set projection is self-adjoint.
Applying a measurable-set projection twice is the same as applying it once.
A measurable projection has the same matrix element when applied in both arguments.
Each measurable-set value of a PVM is an orthogonal (star) projection.
For each measurable set B, the operator P(B) is positive.
Measurable-set projections commute, because intersection is symmetric.
Projections of disjoint measurable sets have zero product.
For measurable B ⊆ C, the projection product P(B) * P(C) is P(B).
Right absorption: if B ⊆ C, then P(C) * P(B) = P(B).
If measurable B ⊆ C, then the range of P(B) is contained in the
range of P(C).
For disjoint measurable sets, the projection of their union is the sum of their projections.
Complementation: the projection of a complement is the complementary projection.
The squared norm of the projection of ξ equals its diagonal measure
on the measurable set, interpreted as a real number.
Total mass of the diagonal scalar measure, as a real number.
The total mass of the diagonal measure as a real number.
The total mass of the diagonal measure in ℝ≥0∞.
For every measurable set B, P(B) is a contraction on vectors.
Strong countable additivity #
For a pairwise-disjoint measurable family (B_k) with measurable union B,
write S n = ∑_{k < n} P(B_k). The diagonal measures are genuine measures,
so μ(B) = ∑' μ(B_k) for every ξ; the squared distance
‖(S n) ξ - P(B) ξ‖² then equals μ(B) - ∑_{k < n} μ(B_k) — the mass of
B not yet covered — and that tends to 0. This is strong countable
additivity: the partial sums converge strongly to P(B).
For pairwise disjoint measurable sets, the finite sums of their projections, applied to any vector, converge to the projection of their union applied to that vector.
Unconditional strong countable additivity: disjoint measurable-set
projections form a summable family of vectors, and their sum is the projection
of the union. This is the HasSum interface needed by the spectral integral.
Extensionality #
A ProjValMeasure is determined by either its projection field or its
finite diagonal scalar measures. The identity inner_proj relates these
fields; finite measure values are recovered from their real embeddings.
Two projection-valued measures with the same diagonal scalar measures are equal.
Two projection-valued measures agreeing on all measurable-set projections are equal.
Equality of projection-valued measures is equivalent to equality of all diagonal scalar measures.