GPD Moments and Polynomiality Interfaces #
This module defines Mellin moments for GPDs and a polynomiality interface, with concrete low-order consequences.
Conventions #
Integration range. Moments run over the full GPD support
x ∈ [-1, 1](seeEpsilonEridani.Particles.Parton.GPD.Basic). The half-range integral∫_0^1is not a polynomial inξ, so polynomiality cannot be stated with it.Moment index. Following
PDF.mellinMoment, the integrand carriesx ^ n, i.e.mellinMomentH M n i ξ t = ∫_{-1}^{1} dx x^n H_i(x, ξ, t). The literature's n-th moment is∫_{-1}^{1} dx x^{n-1} H(Diehl, Generalized parton distributions, Phys. Rept. 388 (2003) 41, arXiv:hep-ph/0307382 — general reference; the primary source was not consulted directly, so no equation number is cited), so the literature index isn + 1in the notation used here. Two consequences of the shift, both visible inPolynomialityAssumptionsbelow:- the polynomiality degree bound is
n + 1, notn; - the D-term occupies the
ξ^(n+1)coefficient, which is why the sum runs overFinset.range (n + 2)rather thanFinset.range (n + 1).
- the polynomiality degree bound is
Skewness range. Polynomiality is a statement about physical kinematics
|ξ| ≤ 1; outside that range the[-1, 1]integral truncates the support of a model and no polynomial statement survives. The hypothesis is therefore carried explicitly.
Mellin moment of a bare GPD function over the full support x ∈ [-1, 1].
The integrand is x ^ n * H i x ξ t; see the module docstring for the index convention.
Equations
Instances For
Mellin moments of H over the full GPD support x ∈ [-1, 1].
Equations
Instances For
Mellin moments of E over the full GPD support x ∈ [-1, 1].
Equations
Instances For
Polynomiality schema for Mellin moments of H.
For physical skewness |ξ| ≤ 1, the moment ∫_{-1}^{1} dx x^n H is a polynomial in ξ
of degree at most n + 1 containing only even powers. The ξ^(n+1) coefficient is the
D-term contribution and is nonzero only for odd n (equivalently, for even literature
index n + 1).
This bundle is retained for models that are not built from a double distribution. For
models that are, GPD.polynomialityAssumptionsOfDoubleDistribution constructs it — but
that bridge is not a discharge: its polynomial field is
GPD.mellinMomentGpd_ofDoubleDistribution, which is a tagged sorry, so #print axioms
on the bridge reports sorryAx. A reader deciding whether this
assumption is "handled" should read it as: handled modulo one open measure-theoretic
statement, and otherwise assumed. No unconditional discharge of this bundle exists
anywhere in the repository at present.
The coefficient of
ξ^kin then-th moment for flavoriat momentum transfert.Moments are even polynomials in
ξ: odd coefficients vanish.- polynomial (n : ℕ) (i : Flavor) (xi t : ℝ) : |xi| ≤ 1 → mellinMomentH M n i xi t = ∑ k ∈ Finset.range (n + 2), self.coeff n i k t * xi ^ k
The moment is the polynomial with those coefficients, for physical skewness.
Instances For
Polynomiality makes every moment an even function of the skewness.
First nontrivial polynomiality consequence: the zeroth moment is ξ-independent.
With the corrected degree bound the n = 0 moment is A + C ξ, and it is the evenness
clause, not the degree bound, that kills the linear term.