Documentation

TauCeti.Analysis.PDE.Yamabe.Bubble.Moments

Moments of the Euclidean Yamabe bubbles #

Let U = U_{t,b} be a Yamabe bubble on an n-dimensional real inner product space. This file computes the radial moments ∫ ‖x - b‖ ^ p … of order p ≥ 0 of its critical density U ^ pₙ, of its Dirichlet-energy density |dU|², and of its square U². For t > 0,

The exact values hold for every additive Haar measure and retain the beta integral and the measure of the unit ball. This form makes both the scale and the loss of integrability at the critical moment transparent. In particular, the fourth energy moment is covered in dimensions at least seven. The integrability and the value of ∫ U ^ pₙ in TauCeti.Analysis.PDE.Yamabe.Bubble.Energy are the moments of order 0.

At the endpoint orders p = n for U ^ pₙ (in dimension n ≠ 2 with n > 0), and p = n - 2 for |dU|² and p = n - 4 for U² (both in dimension n > 2), the moments diverge logarithmically. For these the file computes the moments truncated to a ball B(b, R) exactly, in the same dimensions. Each is a constant times t ^ q (log (1 + (R / t)²) - S_j), where q is n for U ^ pₙ and n - 2 for the other two densities, and S_j is the partial logarithmic series ∑_{k < j} w ^ (k + 1) / (k + 1) in w = R² / (t² + R²) ∈ [0, 1), with j = n - 1 for U ^ pₙ and |dU|² and j = n - 3 for U². Hence the truncated moment is the logarithmic term up to an error of at most the same constant times t ^ q H_j, with H_j the j-th harmonic number. In dimension six these are the fourth moment of |dU|² and the second moment of U², which grow like t⁴ log (1 / t) as t → 0.

Main declarations #

References #

In dimension n ≠ 2, the moment of order 0 ≤ p < n about the centre of the critical bubble density is integrable for every additive Haar measure.

The moment of order 0 ≤ p < n about the centre of the critical bubble density, in dimension n ≠ 2, for an additive Haar measure μ. Its dependence on the scale is exactly t ^ p.

Every positive critical-mass moment of order less than the dimension tends to zero as the bubble scale tends to zero from above, for any additive Haar measure.

In dimension n > 2, the moment of order 0 ≤ p < n - 2 about the centre of the Dirichlet-energy density of a bubble is integrable for every additive Haar measure.

theorem TauCeti.Yamabe.integral_norm_sub_rpow_mul_norm_fderiv_bubble_sq {E : Type u_1} [NormedAddCommGroup E] [InnerProductSpace ℝ E] [FiniteDimensional ℝ E] [MeasurableSpace E] [BorelSpace E] (μ : MeasureTheory.Measure E) [μ.IsAddHaarMeasure] {p : ℝ} (hp : 0 ≤ p) (hpn : p < ↑(Module.finrank ℝ E) - 2) {t : ℝ} (ht : 0 < t) (b : E) :
∫ (x : E), ‖x - b‖ ^ p * ‖fderiv ℝ (bubble t b) x‖ ^ 2 ∂μ = (↑(Module.finrank ℝ E) - 2) ^ 2 * ↑(Module.finrank ℝ E) * μ.real (Metric.ball 0 1) * (2 ^ (Module.finrank ℝ E - 3) * t ^ p * ProbabilityTheory.beta ((↑(Module.finrank ℝ E) + p + 2) / 2) ((↑(Module.finrank ℝ E) - p - 2) / 2))

The moment of order 0 ≤ p < n - 2 about the centre of the Dirichlet-energy density of a bubble, for an additive Haar measure μ. It scales as t ^ p; the excluded endpoint p = n - 2 is the logarithmic case.

Every positive Dirichlet-energy moment of order less than n - 2 tends to zero as the bubble scale tends to zero from above, for any additive Haar measure.

Finite moments of the square #

theorem TauCeti.Yamabe.integral_mul_norm_sub_rpow_mul_bubble_sq_eq_mul {E : Type u_1} [NormedAddCommGroup E] [InnerProductSpace ℝ E] [FiniteDimensional ℝ E] [MeasurableSpace E] [BorelSpace E] (μ : MeasureTheory.Measure E) [μ.IsAddHaarMeasure] (χ : E → ℝ) (p : ℝ) {t : ℝ} (ht : 0 < t) (b : E) :
∫ (x : E), χ x * (‖x - b‖ ^ p * bubble t b x ^ 2) ∂μ = t ^ (p + 2) * ∫ (z : E), χ (b + t • z) * (‖z‖ ^ p * bubble 1 0 z ^ 2) ∂μ

The square moment against an arbitrary real weight rescales by t ^ (p + 2), with the weight pulled back along z ↦ b + t • z. No integrability or restriction on the moment order or dimension is required for this identity of totalized integrals.

theorem TauCeti.Yamabe.integrable_norm_sub_rpow_mul_bubble_sq {E : Type u_1} [NormedAddCommGroup E] [InnerProductSpace ℝ E] [MeasurableSpace E] [BorelSpace E] (μ : MeasureTheory.Measure E) [μ.IsAddHaarMeasure] {p : ℝ} (hp : 0 ≤ p) (hpn : p < ↑(Module.finrank ℝ E) - 4) {t : ℝ} (ht : 0 < t) (b : E) :
MeasureTheory.Integrable (fun (x : E) => ‖x - b‖ ^ p * bubble t b x ^ 2) μ

The moment of order 0 ≤ p < n - 4 of the square of a bubble is integrable for every additive Haar measure.

theorem TauCeti.Yamabe.integral_norm_sub_rpow_mul_bubble_sq {E : Type u_1} [NormedAddCommGroup E] [InnerProductSpace ℝ E] [MeasurableSpace E] [BorelSpace E] (μ : MeasureTheory.Measure E) [μ.IsAddHaarMeasure] {p : ℝ} (hp : 0 ≤ p) (hpn : p < ↑(Module.finrank ℝ E) - 4) {t : ℝ} (ht : 0 < t) (b : E) :
∫ (x : E), ‖x - b‖ ^ p * bubble t b x ^ 2 ∂μ = ↑(Module.finrank ℝ E) * μ.real (Metric.ball 0 1) * (2 ^ (Module.finrank ℝ E - 3) * t ^ (p + 2) * ProbabilityTheory.beta ((↑(Module.finrank ℝ E) + p) / 2) ((↑(Module.finrank ℝ E) - p - 4) / 2))

The moment of order 0 ≤ p < n - 4 of U² equals n μ(B(0,1)) 2^(n-3) Β((n+p)/2, (n-p-4)/2) t^(p+2). The excluded endpoint p = n - 4 is the logarithmic case.

Truncated moments at the logarithmic endpoint #

The moments of order n of U ^ pₙ (in dimension n ≠ 2 with n > 0), of order n - 2 of |dU|² and of order n - 4 of U² (both in dimension n > 2) diverge logarithmically, and the moments truncated to a ball are computed instead. Each density is a power of the bubble profile 2 t / (t² + ‖x - b‖²), so these are specializations of TauCeti.setIntegral_ball_norm_sub_rpow_mul_div_add_norm_sub_sq_pow.

theorem TauCeti.Yamabe.setIntegral_ball_norm_sub_rpow_mul_bubble_rpow_criticalExponent {E : Type u_1} [NormedAddCommGroup E] [InnerProductSpace ℝ E] [FiniteDimensional ℝ E] [MeasurableSpace E] [BorelSpace E] (μ : MeasureTheory.Measure E) [μ.IsAddHaarMeasure] (hn₀ : 0 < Module.finrank ℝ E) (hn : Module.finrank ℝ E ≠ 2) {t : ℝ} (ht : 0 < t) {R : ℝ} (hR : 0 ≤ R) (b : E) :
∫ (x : E) in Metric.ball b R, ‖x - b‖ ^ ↑(Module.finrank ℝ E) * bubble t b x ^ criticalExponent (Module.finrank ℝ E) ∂μ = ↑(Module.finrank ℝ E) * μ.real (Metric.ball 0 1) * (2 ^ (Module.finrank ℝ E - 1) * t ^ Module.finrank ℝ E * (Real.log (1 + (R / t) ^ 2) - ∑ k ∈ Finset.range (Module.finrank ℝ E - 1), (R ^ 2 / (t ^ 2 + R ^ 2)) ^ (k + 1) / (↑k + 1)))

The truncated moment of order n of the critical bubble density, in dimension n ≠ 2, over the ball B(b, R). It grows like n |B(0, 1)| 2 ^ (n - 1) t ^ n log (1 + (R / t)²), and the correction is the partial logarithmic series in w = R² / (t² + R²).

The truncated moment of order n of the critical bubble density, in dimension n ≠ 2, differs from n |B(0, 1)| 2 ^ (n - 1) t ^ n log (1 + (R / t)²) by at most n |B(0, 1)| 2 ^ (n - 1) t ^ n H_{n - 1}.

theorem TauCeti.Yamabe.setIntegral_ball_norm_sub_rpow_mul_norm_fderiv_bubble_sq {E : Type u_1} [NormedAddCommGroup E] [InnerProductSpace ℝ E] [FiniteDimensional ℝ E] [MeasurableSpace E] [BorelSpace E] (μ : MeasureTheory.Measure E) [μ.IsAddHaarMeasure] (hn : 2 < Module.finrank ℝ E) {t : ℝ} (ht : 0 < t) {R : ℝ} (hR : 0 ≤ R) (b : E) :
∫ (x : E) in Metric.ball b R, ‖x - b‖ ^ (↑(Module.finrank ℝ E) - 2) * ‖fderiv ℝ (bubble t b) x‖ ^ 2 ∂μ = (↑(Module.finrank ℝ E) - 2) ^ 2 * ↑(Module.finrank ℝ E) * μ.real (Metric.ball 0 1) * (2 ^ (Module.finrank ℝ E - 3) * t ^ (Module.finrank ℝ E - 2) * (Real.log (1 + (R / t) ^ 2) - ∑ k ∈ Finset.range (Module.finrank ℝ E - 1), (R ^ 2 / (t ^ 2 + R ^ 2)) ^ (k + 1) / (↑k + 1)))

The truncated moment of order n - 2 of the Dirichlet-energy density of a bubble, in dimension n > 2, over the ball B(b, R). It grows like (n - 2)² n |B(0, 1)| 2 ^ (n - 3) t ^ (n - 2) log (1 + (R / t)²); in dimension 6 this is the fourth energy moment.

The truncated moment of order n - 2 of the Dirichlet-energy density of a bubble, in dimension n > 2, differs from (n - 2)² n |B(0, 1)| 2 ^ (n - 3) t ^ (n - 2) log (1 + (R / t)²) by at most (n - 2)² n |B(0, 1)| 2 ^ (n - 3) t ^ (n - 2) H_{n - 1}.

theorem TauCeti.Yamabe.setIntegral_ball_norm_sub_rpow_mul_bubble_sq {E : Type u_1} [NormedAddCommGroup E] [InnerProductSpace ℝ E] [FiniteDimensional ℝ E] [MeasurableSpace E] [BorelSpace E] (μ : MeasureTheory.Measure E) [μ.IsAddHaarMeasure] (hn : 2 < Module.finrank ℝ E) {t : ℝ} (ht : 0 < t) {R : ℝ} (hR : 0 ≤ R) (b : E) :
∫ (x : E) in Metric.ball b R, ‖x - b‖ ^ (↑(Module.finrank ℝ E) - 4) * bubble t b x ^ 2 ∂μ = ↑(Module.finrank ℝ E) * μ.real (Metric.ball 0 1) * (2 ^ (Module.finrank ℝ E - 3) * t ^ (Module.finrank ℝ E - 2) * (Real.log (1 + (R / t) ^ 2) - ∑ k ∈ Finset.range (Module.finrank ℝ E - 3), (R ^ 2 / (t ^ 2 + R ^ 2)) ^ (k + 1) / (↑k + 1)))

The truncated moment of order n - 4 of the square of a bubble, in dimension n > 2, over the ball B(b, R). It grows like n |B(0, 1)| 2 ^ (n - 3) t ^ (n - 2) log (1 + (R / t)²); in dimension 6 this is the second moment of U².

The truncated moment of order n - 4 of the square of a bubble, in dimension n > 2, differs from n |B(0, 1)| 2 ^ (n - 3) t ^ (n - 2) log (1 + (R / t)²) by at most n |B(0, 1)| 2 ^ (n - 3) t ^ (n - 2) H_{n - 3}.

theorem TauCeti.Yamabe.setIntegral_ball_norm_sub_sq_mul_bubble_sq_of_finrank_eq_six {E : Type u_1} [NormedAddCommGroup E] [InnerProductSpace ℝ E] [FiniteDimensional ℝ E] [MeasurableSpace E] [BorelSpace E] (μ : MeasureTheory.Measure E) [μ.IsAddHaarMeasure] (hn : Module.finrank ℝ E = 6) {t : ℝ} (ht : 0 < t) {R : ℝ} (hR : 0 ≤ R) (b : E) :
∫ (x : E) in Metric.ball b R, ‖x - b‖ ^ 2 * bubble t b x ^ 2 ∂μ = 48 * μ.real (Metric.ball 0 1) * t ^ 4 * (Real.log (1 + (R / t) ^ 2) - ∑ k ∈ Finset.range 3, (R ^ 2 / (t ^ 2 + R ^ 2)) ^ (k + 1) / (↑k + 1))

In dimension six, the truncated scalar second moment of a bubble square is an explicit logarithmic term minus a partial logarithmic series.

theorem TauCeti.Yamabe.abs_setIntegral_ball_norm_sub_sq_mul_bubble_sq_sub_mul_log_le_of_finrank_eq_six {E : Type u_1} [NormedAddCommGroup E] [InnerProductSpace ℝ E] [FiniteDimensional ℝ E] [MeasurableSpace E] [BorelSpace E] (μ : MeasureTheory.Measure E) [μ.IsAddHaarMeasure] (hn : Module.finrank ℝ E = 6) {t : ℝ} (ht : 0 < t) {R : ℝ} (hR : 0 ≤ R) (b : E) :
|∫ (x : E) in Metric.ball b R, ‖x - b‖ ^ 2 * bubble t b x ^ 2 ∂μ - 48 * μ.real (Metric.ball 0 1) * t ^ 4 * Real.log (1 + (R / t) ^ 2)| ≤ 48 * μ.real (Metric.ball 0 1) * t ^ 4 * ↑(harmonic 3)

In dimension six, the scalar second moment differs from its logarithmic leading term by at most 48 |B(0,1)| t⁴ H₃.