import OAI.Analysis.Fourier.SquareFunction.Maximal.BRNorm open OAI.Analysis.Fourier.Wave /-! The maximal Bochner--Riesz implication, for the actual supremum over all positive radii. Smooth factors are controlled pointwise by Hardy--Littlewood, before the strict-index shell series is summed. -/ open _root_.MeasureTheory open scoped ENNReal ContDiff noncomputable section namespace OAI.Analysis.Fourier.SquareFunction theorem localSmoothing_maximalBochnerRiesz (hls : OAI.Analysis.Fourier.LocalSmoothing.EuclideanLocalSmoothing) {n : ℕ} (hn : 2 ≤ n) {p δ : ℝ} (hp : 2 ≤ p) (hδ : OAI.Analysis.Fourier.LocalSmoothing.smoothingOrder n p < δ) : ∃ C : ℝ, 0 < C ∧ ∀ f : SchwartzMap (Space n) ℂ, MemLp (OAI.Analysis.Fourier.BochnerRiesz.bochnerRieszMaximal δ f) (ENNReal.ofReal p) volume ∧ eLpNorm (OAI.Analysis.Fourier.BochnerRiesz.bochnerRieszMaximal δ f) (ENNReal.ofReal p) volume ≤ ENNReal.ofReal C * eLpNorm f (ENNReal.ofReal p) volume := by have hδ0 : 0 < δ := lt_of_le_of_lt (le_max_left _ _) hδ have hp' : 1 < p := by linarith have hp1 : 1 ≤ ENNReal.ofReal p := by simpa only [ENNReal.ofReal_one] using ENNReal.ofReal_le_ofReal hp'.le let ε : ℝ := (δ - OAI.Analysis.Fourier.LocalSmoothing.smoothingOrder n p) / 2 have hε : 0 < ε := by dsimp [ε]; linarith have hgap : OAI.Analysis.Fourier.LocalSmoothing.smoothingOrder n p + ε < δ := by dsimp [ε]; linarith obtain ⟨χ, hχ, hc, hzero, hone, _⟩ := OAI.Analysis.Fourier.BochnerRiesz.exists_annular_cutoff (n := n) (a := 1 / 4) (b := 1 / 2) (c := 2) (d := 4) (by norm_num) (by norm_num) (by norm_num) (by norm_num) obtain ⟨C₀, hC₀, L, _, hmax⟩ := OAI.Analysis.Fourier.BochnerRiesz.localSmoothing_maximalShellCycles hls hn (OAI.Analysis.Fourier.BochnerRiesz.dyadicPowerProfile δ) (OAI.Analysis.Fourier.BochnerRiesz.hasCompactSupport_dyadicPowerProfile δ) hp hε obtain ⟨J₀, hJ₀⟩ := pow_unbounded_of_one_lt L (by norm_num : (1 : ℝ) < 2) let J : ℕ := J₀ + 2 have hJ : 2 ≤ J := by dsimp [J]; omega have hlarge (j : ℕ) : L ≤ (2 : ℝ) ^ (J + j) := hJ₀.le.trans (pow_le_pow_right₀ (by norm_num) (by dsimp [J]; omega)) obtain ⟨K, hK, hpacket⟩ := brScaledPacket_maximal_bound (δ := δ) hχ hc hzero obtain ⟨B, hB, hinterior⟩ := brScaledInterior_maximal_bound (n := n) hδ0 hJ obtain ⟨D, hD, hHL⟩ := exists_ballMaximal_envelope_constant n hp' let G : ℝ := (1 - (2 : ℝ) ^ ((OAI.Analysis.Fourier.LocalSmoothing.smoothingOrder n p + ε) - δ))⁻¹ have hG : 0 < G := OAI.Analysis.Fourier.BochnerRiesz.dyadic_rpow_sum_constant_pos hgap let C : ℝ := B * D + (K * D * C₀) * G have hC : 0 < C := by dsimp [C]; positivity refine ⟨C, hC, ?_⟩ intro f have hi := eLpNorm_brMaximalInteriorEnvelope_le hp' hB.le hHL f have hs := eLpNorm_brMaximalShellEnvelope_le hp' hK.le hC₀.le hD.le hgap J hHL f (fun j => (hmax _ (hlarge j) f).2) have H : eLpNorm (OAI.Analysis.Fourier.BochnerRiesz.bochnerRieszMaximal δ f) (ENNReal.ofReal p) volume ≤ ENNReal.ofReal C * eLpNorm f (ENNReal.ofReal p) volume := by calc _ ≤ eLpNorm (fun x => ENNReal.ofReal B * ballMaximal volume (fun y => ‖f y‖ₑ) x + brMaximalShellEnvelope δ K J f x) (ENNReal.ofReal p) volume := by apply eLpNorm_mono_enorm (measurable_bochnerRieszMaximal hδ0.le f).aestronglyMeasurable intro x simp only [enorm_eq_self] exact bochnerRieszMaximal_le_shell_envelope hδ0 hJ hχ hc hzero hone hinterior hpacket f x _ ≤ eLpNorm (fun x => ENNReal.ofReal B * ballMaximal volume (fun y => ‖f y‖ₑ) x) (ENNReal.ofReal p) volume + eLpNorm (brMaximalShellEnvelope δ K J f) (ENNReal.ofReal p) volume := eLpNorm_add_le hp1 _ ≤ ENNReal.ofReal (B * D) * eLpNorm f (ENNReal.ofReal p) volume + ENNReal.ofReal ((K * D * C₀) * G) * eLpNorm f (ENNReal.ofReal p) volume := add_le_add hi hs _ = _ := by rw [← add_mul, ← ENNReal.ofReal_add (by positivity) (by positivity)] exact ⟨H.trans_lt (ENNReal.mul_lt_top ENNReal.ofReal_lt_top (f.memLp _ volume).eLpNorm_lt_top), H⟩ theorem localSmoothing_maximalBochnerRieszConjecture (hls : OAI.Analysis.Fourier.LocalSmoothing.EuclideanLocalSmoothing) : OAI.Analysis.Fourier.BochnerRiesz.MaximalBochnerRieszConjecture := by intro n hn p δ hp hδ exact localSmoothing_maximalBochnerRiesz hls hn hp hδ end OAI.Analysis.Fourier.SquareFunction