Large-data L 2 L^{2} -decay for attractive-dissipative nonlinear Schrödinger equations without the strong dissipative condition

Naoyasu Kita, Hayato Miyazaki, Takuya Sato · 2026

Read Large-data L 2 L^{2} -decay for attractive-dissipative nonlinear Schrödinger equations without the strong dissipative condition by Naoyasu Kita, Hayato Miyazaki, Takuya Sato (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-2605-15837.txt.

Introduction

Naoyasu Kita, Hayato Miyazaki and Takuya Sato prove a large-data L2-decay estimate for attractive-dissipative nonlinear Schrödinger equations. The energy lacks a fixed sign when the real part of lambda is negative, blocking the uniform gradient bound. They add a multiple of the shrinking L2-norm to form an augmented energy, which adds a dissipative term and yields a uniform-in-time H1 bound. This extends the known decay rate across the sharp range 1 < p ≤ 1 + 2/d for arbitrary initial data, removing the last restriction in this case.

Catalog record

CREATORNaoyasu Kita, Hayato Miyazaki, Takuya Sato
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE26188
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.15837

Browse the full library · About the Tower of Babel

Full text

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

Large-data L 2 L^{2} -decay for attractive-dissipative nonlinear Schrödinger equations without the strong dissipative condition Naoyasu Kita, Hayato Miyazaki, Takuya Sato Abstract . We prove a large-data L 2 L^{2} -decay estimate for nonlinear dissipative Schrödinger equations with attractive-dissipative power nonlinearity. The main difficulty is the lack of sign definiteness of the standard energy when Re ⁡ λ < 0 \operatorname{Re}\lambda<0 , which prevents the usual energy argument from directly yielding a uniform gradient bound. We introduce an augmented energy, obtained by adding a suitable multiple of the decreasing L 2 L^{2} -norm to the standard energy. This produces an additional dissipative term and gives a direct uniform-in-time H 1 H^{1} bound without the iteration argument used in previous works. Consequently, for arbitrary initial data in the weighted energy space Σ = H 1 ∩ ℱ ​ H 1 \Sigma=H^{1}\cap\mathcal{F}H^{1} , we obtain the decay rate previously known under the strong dissipative condition throughout the sharp decay range 1 < p ⩽ 1 + 2 / d 1<p\leqslant 1+2/d . This removes the remaining restriction p ⩽ 1 + 4 / ( 3 ​ d ) p\leqslant 1+4/(3d) in the attractive-dissipative case. 2020 Mathematics Subject Classification 35Q55, 35B40 1. Introduction We consider the nonlinear Schrödinger equation (NLS) \displaystyle\begin{cases}\displaystyle i\partial_{t}u+\frac{1}{2}\Delta u=\lambda|u|^{p-1}u,\quad(t,x)\in(0,\infty)\times\mathbb{R}^{d},\\ u(0,x)=u_{0}(x),\quad x\in\mathbb{R}^{d},\end{cases} where d\in\mathbb{N} , p>1 , and \lambda\in\mathbb{C} satisfies \operatorname{Im}\lambda<0. Then the L^{2} -norm is dissipative: \displaystyle\lVert u(t)\rVert_{2}^{2}=\lVert u_{0}\rVert_{2}^{2}+2\operatorname{Im}\lambda\int_{0}^{t}\lVert u(s)\rVert_{p+1}^{p+1}\,ds. The purpose of this paper is to establish the sharp large-data decay rate of \lVert u(t)\rVert_{2} in the attractive-dissipative case \operatorname{Re}\lambda<0,\qquad\operatorname{Im}\lambda<0. For general solutions, the natural range for the L^{2} -decay problem is 1<p\leqslant 1+\frac{2}{d}. Indeed, when p>1+2/d , the nonlinear dissipation is too weak in general, and the L^{2} -norm need not decay to zero; see, for instance, [ 19 ] . Under the strong dissipative condition (1.1) \displaystyle(p-1)|\lambda|\leqslant(p+1)|\operatorname{Im}\lambda|,\qquad\operatorname{Im}\lambda<0, Hayashi, Li, and Naumkin [ 11 ] proved that arbitrary data u_{0}\in\Sigma:=H^{1}(\mathbb{R}^{d})\cap\mathcal{F}H^{1}(\mathbb{R}^{d}) give the decay rate (1.2) \displaystyle\lVert u(t)\rVert_{2}\lesssim\begin{cases}(\log t)^{-\frac{2}{(d+2)(p-1)}}&\text{if }p=1+2/d,\\ t^{-\frac{2}{(d+2)(p-1)}\left(1-\frac{d(p-1)}{2}\right)}&\text{if }1<p<1+2/d.\end{cases} The difficulty in removing ( 1.1 ) depends on the sign of \operatorname{Re}\lambda . When \operatorname{Re}\lambda\geqslant 0 , the standard energy gives a uniform bound for \lVert\nabla u(t)\rVert_{2} , and the decay rate ( 1.2 ) was obtained in [ 8 ] . When \operatorname{Re}\lambda<0 , however, the standard energy E(u)=\frac{1}{2}\lVert\nabla u\rVert_{2}^{2}+\frac{\operatorname{Re}\lambda}{p+1}\lVert u\rVert_{p+1}^{p+1} is not sign definite. Thus the standard energy does not directly control \lVert\nabla u(t)\rVert_{2} . Recent works [ 9 ] and [ 17 ] overcame this difficulty only in restricted ranges of p ; in particular, [ 17 ] reached p\leqslant 1+\frac{4}{3d}. Thus the remaining range 1+\frac{4}{3d}<p\leqslant 1+\frac{2}{d} was left open in the attractive-dissipative case. Dissipative nonlinear Schrödinger equations have also been studied in connection with refined asymptotic behavior and pointwise decay; see [ 4 , 3 , 6 , 5 , 15 , 14 , 12 , 13 ] . These works often impose smoothness, spatial decay, or additional structural conditions on the initial data. In contrast, we prove an L^{2} -decay estimate for arbitrary data in \Sigma . Another related direction concerns nonlinear Schrödinger equations with combined focusing and dissipative nonlinearities, where finite-time blow-up and global existence have been studied; see [ 1 , 2 , 7 ] . The present paper treats a single-power attractive-dissipative nonlinearity and focuses on the quantitative L^{2} -decay of global solutions in the range 1<p\leqslant 1+2/d . The main observation of this paper is that the mass dissipation itself can be used to restore a coercive energy control. For a suitable constant \gamma>0 , we introduce the augmented energy E_{\mathrm{aug}}(u)=E(u)+\gamma\lVert u\rVert_{2}^{2}. The time derivative of the added term produces the negative contribution -2\gamma|\operatorname{Im}\lambda|\lVert u\rVert_{p+1}^{p+1}, which absorbs the noncoercive part of the standard energy estimate. This bypasses the bootstrap used in earlier approaches, where intermediate L^{2} -decay estimates are substituted into energy estimates, and yields a direct uniform-in-time bound for \lVert\nabla u(t)\rVert_{2} . Consequently, we obtain ( 1.2 ) throughout the whole range 1<p\leqslant 1+2/d . We use the standard integral formulation of solutions. Let I\subset\mathbb{R} be an interval and fix t_{0}\in I . A function u\in C(I;L^{2}(\mathbb{R}^{d})) is called a solution to ( NLS ) on I if u(t)=U(t-t_{0})u(t_{0})-i\int_{t_{0}}^{t}U(t-s)\lambda|u(s)|^{p-1}u(s)\,ds in L^{2}(\mathbb{R}^{d}) for all t,t_{0}\in I , where U(t)=e^{it\Delta/2} . We now state the main result. Theorem 1.1 . Let 1<p\leqslant 1+2/d . Assume \operatorname{Re}\lambda<0 and \operatorname{Im}\lambda<0 . Take u_{0}\in\Sigma with \lVert u_{0}\rVert_{\Sigma}\neq 0 . Then ( NLS ) admits a unique global solution u\in C([0,\infty);\Sigma) . Moreover, if p=1+2/d , then it holds that \lVert u(t)\rVert_{2}\lesssim(\log(1+t))^{-\frac{2}{(d+2)(p-1)}} for any t\geqslant 1 . When 1<p<1+2/d , the estimate \lVert u(t)\rVert_{2}\lesssim(1+t)^{-\frac{2}{(d+2)(p-1)}\left(1-\frac{d(p-1)}{2}\right)} is valid for any t\geqslant 0 . Remark 1.2 . The argument also applies to the repulsive-dissipative case \operatorname{Re}\lambda\geqslant 0 , \operatorname{Im}\lambda<0 . In that case, the standard energy already gives a uniform gradient bound, and the augmented energy argument recovers the same decay rate without imposing ( 1.1 ). 1.1. Strategy of the proof The proof is based on two elementary estimates and one augmented energy argument. First, the L^{2} -dissipative identity and Hölder’s inequality imply (1.3) \displaystyle\frac{d}{dt}\lVert u(t)\rVert_{2}^{2}\leqslant-C\lVert u(t)\rVert_{2}^{\beta}\lVert u(t)\rVert_{q}^{-(\beta-p-1)} for some \beta>p+1 , where q=1 if d=1 and q=2(d+1)/(d+2) if d\geqslant 2 . Second, the weighted interpolation inequality \lVert u\rVert_{q}\lesssim\lVert u\rVert_{2}^{1-d\alpha}\lVert xu\rVert_{2}^{d\alpha},\qquad\alpha=\frac{1}{q}-\frac{1}{2}, reduces the decay estimate to the growth control of \lVert xu(t)\rVert_{2} . The virial-type estimate \lVert xu(t)\rVert_{2}\lesssim\lVert xu_{0}\rVert_{2}+\int_{0}^{t}\lVert\nabla u(s)\rVert_{2}\,ds shows that it is enough to obtain a uniform-in-time bound for \lVert\nabla u(t)\rVert_{2} . This is precisely where the attractive-dissipative case is difficult, because the standard energy is not sign definite. In previous approaches to the attractive-dissipative case, one had to improve the growth bound for \lVert\nabla u(t)\rVert_{2} through an iteration argument. Here the iteration means that a preliminary growth bound is converted through ( 1.3 ) into an L^{2} -decay estimate, which is then substituted into the energy estimate to improve the growth bound. The point of the present paper is that this iteration can be avoided. We add a suitable multiple of the decreasing L^{2} -norm to the standard energy and prove in Proposition 3.1 that E_{\mathrm{aug}}(u(t))=E(u(t))+\gamma\lVert u(t)\rVert_{2}^{2} is nonincreasing for an appropriate choice of \gamma>0 depending on the size of the initial data. This directly gives a uniform gradient bound, hence \lVert xu(t)\rVert_{2}\lesssim 1+t, and the decay rate ( 1.2 ) follows from ( 1.3 ). Notations For any p\geqslant 1 , L^{p}=L^{p}(\mathbb{R}^{d}) denotes the usual Lebesgue space on \mathbb{R}^{d} equipped with the norm \lVert\,\cdot\,\rVert_{p}\coloneqq\lVert\,\cdot\,\rVert_{L^{p}} . Set \langle a\rangle=(1+|a|^{2})^{1/2} for a\in\mathbb{C} or \mathbb{R}^{d} . \mathcal{F}[u]=\widehat{u} is the usual Fourier transform of a function u on \mathbb{R}^{d} . For s\geqslant 0 , the standard Sobolev space on \mathbb{R}^{d} is defined by H^{s}=H^{s}(\mathbb{R}^{d})\coloneqq\{u\in L^{2}(\mathbb{R}^{d})\mid\lVert u\rVert_{H^{s}}\coloneqq\lVert\langle\nabla\rangle^{s}u\rVert_{L^{2}}<\infty\} . We denote the weighted Sobolev space by \mathcal{F}H^{s}=\mathcal{F}H^{s}(\mathbb{R}^{d})\coloneqq\{u\in L^{2}(\mathbb{R}^{d})\mid\lVert u\rVert_{\mathcal{F}H^{s}}\coloneqq\lVert\langle x\rangle^{s}u\rVert_{L^{2}}<\infty\} . Set \Sigma=H^{1}\cap\mathcal{F}H^{1} . Let U(t) be the Schrödinger group e^{it\Delta/2} . A\lesssim B denotes A\leqslant CB for some constant C>0 , which may change from line to line. The paper is organized as follows. Section 2 collects the preliminary estimates needed for the decay argument. Section 3 introduces the augmented energy method, derives the uniform gradient bound without iteration, and proves Theorem 1.1 . 2. Preliminary In this section, we introduce several preliminary results that will be used in the proof of the main theorem. Let us first recall Strichartz estimates. A pair (p,r)

Read the complete text