Local smoothing estimates for bilinear Fourier integral operators

Duván Cardona 1,∗ · 2026

Read Local smoothing estimates for bilinear Fourier integral operators by Duván Cardona 1,∗ (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-15667.txt.

Introduction

Duván Cardona works on smoothing estimates for Fourier integral operators in a 2026 arXiv paper. The linear local smoothing conjecture is a celebrated open idea from Sogge. It concerns how waves concentrate and decay. Cardona formulates a bilinear version of that conjecture. Bilinear operators handle pairs of functions at once. His version is stated for every dimension two and above.

He then proves a key implication. The linear conjecture, if true, implies his bilinear one. Such links matter in harmonic analysis. A single hard conjecture can unlock a family of results. Smoothing estimates govern how solutions of wave equations behave. They feed into broader questions about waves in complex media. The paper is aimed at specialists in the field.

Catalog record

CREATORDuván Cardona 1,∗
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE76 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2601-15667

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Local smoothing estimates for bilinear Fourier integral operators Thanks: 1 Website: https://sites.google.com/site/duvancardonas/home Duván Cardona is supported by the FWO Odysseus 1 grant G.0H94.18N: Analysis and Partial Differential Equations and by the Methusalem programme of the Ghent University Special Research Fund (BOF) (Grant number 01M01021). Duván Cardona has been supported by the FWO Fellowship Grant No 1204824N and by the FWO Grant K183725N of the Belgian Research Foundation FWO. ∗ Orcid:0000-0002-8148-2624 Duván Cardona 1,∗ Abstract . We formulate a local smoothing conjecture for bilinear Fourier integral operators in every dimension d ≥ 2 , d\geq 2, derived from the celebrated linear case due to Sogge, which we refer to as the bilinear smoothing conjecture . We show that the linear local smoothing conjecture implies this bilinear version. As a consequence of our approach and due to the recent progress on the subject, we establish local smoothing estimates for Fourier integral operators in dimension d = 2 , d=2, that is, on ℝ x 2 × ℝ t \mathbb{R}^{2}_{x}\times\mathbb{R}_{t} . Also, a partial progress is presented for the high-dimensional case d ≥ 3 . d\geq 3. In particular, our method allows us to deduce that the bilinear local smoothing conjecture holds for all odd dimensions d d . Keywords. Local smoothing conjecture, bilinear Fourier integral operators, bilinear smoothing conjecture, cinematic curvature condition, wave equation. 2020 Mathematics Subject Classification. Primary 47A10; Secondary 34L20. 1. Introduction This note proposes a conjecture on local smoothing estimates for bilinear Fourier integral operators (FIOs) derived from the linear case observed by Sogge [ 52 ] . Recent progress in the field (see [ 4 , 26 ] ) allows our methods to fully prove the conjecture for d=2 and to advance the problem partially for d\geq 3 . In particular, we deduce that the bilinear local smoothing conjecture is true when d is odd. We first recall the linear local smoothing conjecture [ 52 ] and then introduce its bilinear form. Our main result is Theorem 2.6 ; our approach investigates the smoothing properties of high and of low frequencies of a bilinear Fourier integral operator. For the analysis of the high frequency portion, among other things, we make use of a boundedness theorem for maximal functions proved by Bourgain in [ 5 ] . According to Hörmander [ 30 ] , a Fourier integral operator T of order m\in\mathbb{R} and with real phase function \phi, is a continuous linear operator T:C^{\infty}_{0}(\mathbb{R}^{d})\rightarrow C^{\infty}(\mathbb{R}^{d}\times\mathbb{R}) of the form T_{a}^{\phi}f(x,t)=\smallint\limits_{\mathbb{R}^{n}}e^{i\phi(x,t,\xi)}a(x,t,\xi)\widehat{f}(\xi)d\xi, (1.1) where \widehat{f} is the Fourier transform of f\in C^{\infty}_{0}(\mathbb{R}^{d}). The symbol a:=a(x,t,\xi) is of Kohn-Nirenberg order m\in\mathbb{R}^{d}, that is a satisfies the symbol inequalities \forall\alpha\in\mathbb{N}_{0},\,\forall\beta\in\mathbb{N}_{0},\,\,|\partial_{x}^{\beta}\partial_{\xi}^{\alpha}a(x,t,\xi)|\leq C_{\alpha,\beta}(1+|\xi|)^{m-|\alpha|}. (1.2) The symbol a has compact support in (x,t). We assume that in an open neighbourhood V of \pi(\textnormal{supp}(a)):=\{(x,t):(x,t,\xi)\in\textnormal{supp}(a)\}, (with \overline{V} being a compact set) the real-valued phase function \phi is homogeneous of degree 1 in \xi\neq 0, smooth in (x,t,\xi)\in{V}\times(\mathbb{R}^{d}\setminus\{0\}), and satisfies the non-degeneracy condition \forall(x,t,\xi)\in\overline{V}\times(\mathbb{R}^{d}\setminus\{0\}),\,\textnormal{det}\left((\partial_{x,\xi}^{2}\phi(x,t,\eta))\right)_{1\leq i,j\leq d}\neq 0. (1.3) The corresponding L^{p} -theory for these operators can be summarised as follows, see Tao [ 54 ] and Seeger, Sogge and Stein [ 51 ] . Theorem 1.1 . Let t\in\mathbb{R} be fixed in the support of the symbol a:=a(x,t,\xi) . Let T=T_{a(\cdot,t,\cdot)}^{\phi(\cdot,t,\cdot)} be a Fourier integral operator of order m\in\mathbb{R} . Then: 1. If m=-(d-1)/2, T is of weak (1,1) type: \exists C_{t}>0,\forall f\in L^{1}(\mathbb{R}^{d}),\,\,\|Tf\|_{L^{1,\infty}(\mathbb{R}^{d}_{x})}\leq C_{t}\|f\|_{L^{1}(\mathbb{R}^{d}_{x})}. (1.4) 2. Let 1<p<\infty. Let T be a Fourier integral operator of order m=-(d-1)|\frac{1}{2}-\frac{1}{p}|. Then T is bounded from L^{p}(\mathbb{R}^{d}_{x}) to L^{p}(\mathbb{R}^{d}_{x}). When p=1, then T is bounded from the Hardy space H^{1}(\mathbb{R}^{d}_{x}) to L^{1}(\mathbb{R}^{d}_{x}). As in ( 1.4 ), these operator norms depend on t. The case p=2 in Theorem 1.1 was proved by Duistermaat and Hörmander [ 21 ] , see also Èskin [ 22 ] . Other references related to the global boundedness of Fourier integral operators are presented in Section 5 . For now, we emphasise that, in terms of the L^{p} -regularity, Theorem 1.1 shows that Fourier integral operators lose s=(d-1)|\frac{1}{2}-\frac{1}{p}| derivatives. In other words, on the scale of Sobolev spaces L^{p}_{s}, a Fourier integral operator T=T_{a(\cdot,t,\cdot)}^{\phi(\cdot,t,\cdot)} of order m\in\mathbb{R}, satisfies the time-dependent estimate \|Tf\|_{L^{p}_{s}(\mathbb{R}^{d}_{x})}\lesssim_{t}\|f\|_{L^{p}_{s+m+(d-1)|\frac{1}{2}-\frac{1}{p}|}(\mathbb{R}^{d}_{x})},\,1<p<\infty. (1.5) It was first observed by Sogge [ 52 ] , in the context of the wave equation, that averaging in time on the left-hand side of ( 1.5 ) allows one to improve the loss of regularity. This observation led to the formulation of the (linear) local smoothing conjecture for the wave operator. If true, this conjecture would represent a remarkable breakthrough, as it formally implies several other major open problems in harmonic analysis, including the Bochner–Riesz, Fourier restriction, and Kakeya conjectures, see Tao [ 53 ] . In the setting of Fourier integral operators, the main assumption on the phase to formulate the linear local smoothing conjecture is the following cinematic curvature condition . Definition 1.2 (Cinematic curvature condition) . We say that \phi:=\phi(x,t,\xi)\in C^{\infty}(\mathbb{R}^{d}_{x}\times\mathbb{R}_{t}\times(\mathbb{R}^{d}\setminus\{0\})) satisfies the cinematic curvature condition on a compact set \mathcal{K}\subset\mathbb{R}^{d}\times\mathbb{R}_{t} if the following conditions are satisfied: • (i) the phase \phi is non-degenerate, that is \textnormal{rank}(\partial_{x,\eta}^{2}\phi(x,t,\eta))=d,\quad(x,t)\in\mathcal{K},\quad\eta\in\mathbb{R}^{d}\setminus\{0\}. (1.6) • (ii) Consider the Gauss mapping G:\mathcal{K}\times(\mathbb{R}^{d}\setminus\{0\})\rightarrow\mathbb{S}^{d}, where G(x,t,\eta)=\frac{G_{0}(x,t,\eta)}{|G_{0}(x,t,\eta)|}, and G_{0}(x,t,\eta):=\wedge_{1\leq j\leq d}\partial_{\eta_{j}}\partial_{x,t}\phi(x,t,\eta). The following rank condition \textnormal{rank}(\partial_{\eta\eta}^{2}\langle\partial_{x,t}\phi(x,t,\eta),G(z,\eta_{0})\rangle|_{\eta=\eta_{0}})=d-1, (1.7) holds for every (x,t,\eta_{0})\in\mathcal{K}\times\mathbb{R}^{d}, \eta_{0}\neq 0. Adapted to Fourier integral operators, the linear smoothing conjecture can be stated as follows. In order to formulate it we use the following notation for its critical index: \overline{p}_{d}:=\frac{2(d+1)}{d-1},\textnormal{ if }\,d\textnormal{ is odd, and, }\overline{p}_{d}:=\frac{2(d+2)}{d},\textnormal{ if }\,d\textnormal{ is even}. (1.8) Conjecture 1.3 (Linear smoothing conjecture for FIOs) . Let d\geq 2. Let T_{a}^{\phi} be a Fourier integral operator of order m\in\mathbb{R}, and with non-degenerate phase function satisfying the cinematic curvature condition , see Definition 1.2 . For any p\geq\overline{p}_{d}, and \sigma<\frac{1}{p}, we have that \|T_{a}^{\phi}f\|_{L^{p}(\mathbb{R}^{d}\times\mathcal{I})}\lesssim\|f\|_{L^{p}_{s}(\mathbb{R}^{d})},\,s=m+(d-1)\left(\frac{1}{2}-\frac{1}{p_{i}}\right)-\sigma. (1.9) Remark 1.4 . Examples constructed by Bourgain [ 7 , 8 ] can be used to identify Fourier integral operators that satisfy the cinematic curvature condition but for which local smoothing fails for all \sigma<\frac{1}{p}, whenever p<\overline{p}_{d} . Compared with Theorem 1.1 , Conjecture 1.3 claims a gain of regularity of order \sigma<\frac{1}{p}. We illustrate some values of the critical index \overline{p}_{d} in Table 1 . d 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 \overline{p}_{d} 4 4 3 3 \frac{8}{3} \frac{8}{3} \frac{5}{2} \frac{5}{2} \frac{12}{5} \frac{12}{5} \frac{7}{3} \frac{7}{3} \frac{16}{7} \frac{16}{7} \frac{9}{4} \frac{9}{4} \frac{20}{9} \frac{20}{9} \frac{11}{5} Table 1. Critical index \overline{p}_{d} for dimensions d=2 to 20 . Let us introduce the required notation to formulate the bilinear counterpart to Conjecture 1.3 by adopting the notion of bilinear Fourier integral operators as studied e.g. by Grafakos and Peloso [ 27 ] , and for the form of the operators formulated here, see also Rodríguez-López, Rule, and Staubach [ 33 ] . Let a\in C^{\infty}(\mathbb{R}^{d}_{x}\times\mathbb{R}_{t}\times\mathbb{R}^{d}_{\xi}\times\mathbb{R}^{d}_{\eta}) be a bilinear symbol . We say that a has order m\in\mathbb{R}, if it satisfies the following estimates |\partial_{x,t}^{\beta}\partial_{\xi}^{\alpha_{1}}\partial_{\eta}^{\alpha_{2}}a(x,t,\xi,\eta)|\lesssim_{\beta,\alpha_{1},\alpha_{2}}(1+|\xi|+|\eta|)^{m-|\alpha_{1}|-|\alpha_{2}|}. (1.10) The Kohn-Nirenberg bilinear class S^{m}_{1,0}:=S^{m}_{1,0}(\mathbb{R}^{d}_{x}\times\mathbb{R}_{t}\times\mathbb{R}^{d}_{\xi}\times\mathbb{R}^{d}_{\eta}) is the set formed by those bilinear symbols of order m. Let us consider the bilinear Fourier integral operator \displaystyle T^{\phi_{1},\phi_{2}}_{a}(f,g)(x,t)=\smallint_{\mathbb{R}^{d}}\smallint_{\mathbb{R}^{d}}e^{2\pi i(\phi_{1}(x,t,\xi)+\phi_{2}(x,t,\eta))}a(x,t,\xi,\eta)\widehat{f}(\xi)\widehat{g}(\eta)d\xi d\eta, (1.11) where the symbol a is compactly supported in (x,t)\in\mathbb{R}^{d}\times\mathbb{R}, and each phase function \phi_{j}(x,t,\xi)\in C^{\infty}(\mathbb{R}^{d}_{x}\times\mathbb{R}_{t}\times(\mathbb{R}^{d}\setminus\{0\})) is positively homogeneous of order 1 in \xi\in\mathbb{R}^{d}\setminus\{0\}, and they satisfy the cinematic

Read the complete text