Strichartz and local smoothing estimates for the fractional Schrödinger equations over fractal time
Jin Bong Lee, Sanghyuk Lee, Luz Roncal · 2025
Read Strichartz and local smoothing estimates for the fractional Schrödinger equations over fractal time by Jin Bong Lee, Sanghyuk Lee, Luz Roncal (2025) 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-2508-16102.txt.
Introduction
Jin Bong Lee, Sanghyuk Lee, and Luz Roncal prove estimates for a wave equation in twenty twenty-five. They study the fractional Schrödinger operator, but only over fractal time sets. A fractal set of times has gaps at every scale, unlike a normal interval. Classical Strichartz estimates assume regular time.
To capture the fractal nature, they use tools in the spirit of Assouad dimension. Bounded Assouad characteristic and the Assouad spectrum enter the proofs. They also prove bounds against measures with a set growth condition. Inhomogeneous and local smoothing estimates follow in the same frame. Surprisingly, the known estimates extend in a natural way, matching the fractal dimensions involved. Readers in harmonic analysis get a calculus that survives rough time.
Catalog record
| CREATOR | Jin Bong Lee, Sanghyuk Lee, Luz Roncal |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 93 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2508-16102
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Strichartz and local smoothing estimates for the fractional Schrödinger equations over fractal time Jin Bong Lee, Sanghyuk Lee, Luz Roncal Abstract . We obtain Strichartz-type estimates for the fractional Schrödinger operator f ↦ e i t ( − Δ ) γ / 2 f f\mapsto e^{it(-\Delta)^{\gamma/2}}f over a time set E E of fractal dimension. To obtain those estimates capturing fractal nature of E E , we employ the notions in the spirit of the Assouad dimension, such as, bounded Assouad characteristic and Assouad specturm. We also prove the estimate ‖ e i t ( − Δ ) γ / 2 f ‖ L t q ( d μ , L x r ( ℝ d ) ) ≤ C ‖ f ‖ H s , \|e^{it(-\Delta)^{\gamma/2}}f\|_{L_{t}^{q}(\mathrm{d}\mu;L_{x}^{r}(\mathbb{R}^{d}))}\leq C\|f\|_{H^{s}}, where μ \mu is a measure satisfying an α \alpha -dimensional growth condition. In addition, we establish related inhomogeneous estimates and L 2 L^{2} local smoothing estimates. A surprising feature of our work is that, despite dealing with rough fractal sets, we extend the known estimates for the fractional Schrödinger operators in a natural way, precisely consistent with the associated fractal dimensions. 2020 Mathematics Subject Classification 35J10 (primary); 42B20; 28A80 1. introduction For a positive number \gamma , we consider the operator U^{\gamma}_{t}f(x)=\mathrm{e}^{it(-\Delta)^{\gamma/2}}f(x):=(2\pi)^{-d}\int_{\mathbb{R}^{d}}\mathrm{e}^{i(\langle x,\xi\rangle+t|\xi|^{\gamma})}\widehat{f}(\xi)\,\mathrm{d}\xi, which is a solution to the initial value problem i\partial_{t}u+(-\Delta)^{\gamma/2}u=0,\quad u(0)=u_{0}. The case \gamma=2 corresponds to the classical Schrödinger equation, which arises from the Brownian motion. For \gamma\in(0,2)\setminus\{1\} , the fractional Schrödinger equation is derived by applying Feynman’s path integral framework to Lévy motion; see Laskin [ 18 , 19 ] . Space-time estimates, known as Strichartz estimates, play an important role in the study of both linear and nonlinear equations by providing quantitative control of the solutions. These estimates are formulated in mixed norms in x and t , whose exponents are required to satisfy an admissible condition. We say that a pair (q,r)\in\mathbb{[}2,\infty]\times[2,\infty) is \sigma -admissible if \frac{1}{q}\leq\sigma\Big(\frac{1}{2}-\frac{1}{r}\Big). It is well known that \|\mathrm{e}^{it(-\Delta)^{\gamma/2}}f\|_{L_{t}^{q}(\mathbb{R};L^{r}_{x}(\mathbb{R}^{d}))}\leq C\|f\|_{\dot{H}^{s}},\quad s=\frac{d}{2}-\frac{d}{r}-\frac{\gamma}{q} holds for sharp d/2 -admissible (q,r) , provided that \gamma\in\mathbb{R}_{>0}\setminus\{1\} (see, for example, [ 6 ] ). Here, \dot{H}^{s} denotes the homogeneous Sobolev space. For the wave operator \mathrm{e}^{it(-\Delta)^{1/2}} , the estimate holds for \frac{d-1}{2} -admissible (q,r) . We recommend [ 15 , 6 , 5 , 11 ] and references therein for the previous studies on the Strichartz estimate for the fractional Schrödinger equation. Let E\subset\mathbb{R} be a bounded set. In this note, we are concerned with the Strichartz type estimates over \mathbb{R}^{d}\times E , where the time variable is restricted (from \mathbb{R} ) to a subset E of fractal dimension, and the Lebesgue measure \mathrm{d}t is replaced by a general measure \mu supported on E satisfying certain conditions. This approach enables us to investigate the temporal behavior of the solution on the lower-dimensional set E . Most typical examples of such sets are the Cantor sets. Those kinds of estimates have recently appeared in the works [ 1 , 3 , 24 , 20 ] (see ( 2.8 ) below), where spherical maximal functions defined via the supremum over a fractal set of dilation parameters were studied. More precisely, Strichartz-type estimates for the wave operator were obtained, relating the size (or dimension) of E to the smoothness of the input function. Let N(E,\delta) denote the minimal number of intervals of length \delta needed to cover E . The (upper) Minkowski dimension \dim_{M}E of E is given by the infimum of \beta>0 such that N(E,\delta)\leq C(\varepsilon)\delta^{-\beta-\varepsilon} for every \epsilon>0 . This notion is particularly useful for quantifying the size of a lower-dimensional set E via coverings. However, as shown in the previous works [ 1 , 3 , 24 ] , the Minkowski dimension alone is not enough to capture the dispersive nature of the wave operator. This is due to the multi-scale phenomena associated to frequency localization, i.e., Littlewood–Paley projection. As to be seen later, by using the Minkowski dimension alone, it is difficult to obtain a precise measurement of quantities in relative scale with respect to the size of the frequency of a given function. In this paper, we mainly make use of the Assouad dimension, which provides a more effective framework for obtaining quantitative estimates over fractal sets. The Assouad dimension \dim_{A}E is defined by \dim_{A}E=\inf\{a>0:\exists c>0\text{ s.t. }\forall I,\delta\in(0,|I|),N(E\cap I,\delta)\leq c\delta^{-a}|I|^{a}\}, where I runs over subintervals of [0,1] . It is clear that \dim_{M}E\leq\dim_{A}E . Although the definitions of the Assouad and Minkowski dimensions are similar, the inequality can be strict. For example, consider a set E_{a}=\{n^{-a}:n=1,2,3,\dots\} for 0<a . Then we have \dim_{M}(E_{a})=1/(1+a)<1=\dim_{A}(E_{a}). We note that \dim_{M}(E)=\dim_{A}(E) when E is Ahlfors–David regular (see, for example, [ 8 , Theorem 6.4.1] ). 1.1. Homogeneous estimates In order to get a precise quantitative estimate for the operator U^{\gamma}_{t} over the fractal set E , the Assouad dimension alone is insufficient due to the nature of its definition. Unlike the typical cases, we do not assume that E is contained in a fixed compact set. For \alpha\in(0,1] and a set E\subset\mathbb{R} , we define (1.1) [E]_{\alpha}=\sup_{0<\delta}\sup_{\delta\leq|I|}\Big(\frac{\delta}{|I|}\Big)^{\alpha}N(E\cap I,\delta), where I runs over any interval in \mathbb{R} . This may be regarded as a global version of the notion of bounded \alpha -Assouad characteristic. We say that E has bounded \alpha -Assouad characteristic if [E]_{\alpha}<\infty . Our first result concerns the estimates with frequency localization. Let \psi\in C_{c}^{\infty}(1/4,2) . For j>0 , let P_{j} denote the Littlewood–Paley projection operator defined by (1.2) P_{j}f=(\psi(2^{-j}|\cdot|)\widehat{f}\,)^{\vee},\quad and P_{\leq 0}={\rm Id}-\sum_{j=1}^{\infty}P_{j} . Also, let H^{s} denote the inhomogeneous Sobolev spaces of order s with the norm \|f\|_{H^{s}}^{2}=\int(1+|\xi|^{2})^{s}|\widehat{f}(\xi)|^{2}\,\mathrm{d}\xi. For simplicity, we write L_{x}^{r}:=L^{r}_{x}(\mathbb{R}^{d}) for the remainder of the paper. Theorem 1.1 . Let \alpha\in(0,1] and \gamma\in\mathbb{R}_{>0}\setminus\{1\} , and 2\leq q,r\leq\infty with (d/\alpha,q,r)\neq(2,2,\infty) . Suppose that E is a bounded set with bounded \alpha -Assouad characteristic. For \delta\in\mathbb{R}_{>0} , denote E(\delta)=\{x:\operatorname{dist}(x,E)<\delta\} . Then, the estimate (1.3) \sup_{j\geq 1}\big\|U^{\gamma}_{t}P_{j}f\big\|_{L_{t}^{q}(E(2^{-\gamma j});L_{x}^{r})}+\big\|U^{\gamma}_{t}P_{\leq 0}f\big\|_{L_{t}^{q}(E(1);L_{x}^{r})}\leq C[E]_{\alpha}^{\frac{1}{q}}\|f\|_{H^{s}} holds provided that (q,r) satisfies (1.4) \frac{d}{2}\Big(\frac{1}{2}-\frac{1}{r}\Big)\geq\frac{\alpha}{q} and (1.5) s\geq s_{\gamma}(q,r):=\frac{d}{2}-\frac{d}{r}-\frac{\gamma}{q}. If \dim_{A}E=\alpha , then by a straightforward application of the definition of the Assouad dimension, one obtaians the estimate ( 1.3 ) provided that ( 1.4 ) is replaced by a strict inequality. Using some specific set such as Cantor sets, one can also show that the result is basically sharp. (See Section 2.4 .) When E=\mathbb{R} , so \alpha=1 , the result recovers the classical Strichartz estimate ( [ 14 ] ) via the Littlewood–Paley inequality and scaling. Indeed, once we have the local-in-time estimate, the global-in-time estimate can also be deduced by a scaling argument using homogeneity. As \alpha decreases, the admissible range ( 1.4 ) of (q,r) becomes larger. This is natural, as a smaller \alpha indicates that control is required only on a smaller time set, thus allowing a broader range of admissible pairs. At the extreme case \alpha=0 , for example, when E is a singleton, i.e., E=\{t_{0}\} for some t_{0}\in\mathbb{R} , the estimate ( 1.3 ) holds for all r,q\geq 2 , provided that ( 1.5 ) is satisfied. This is, in fact, a straightforward consequence of the Sobolev embedding. The underlying time set E(2^{-\gamma j}) is also natural in view of the uncertainty principle, since the