Non-Polynomial Conserved Quantities for ODE Systems and its Application to the Long-Time Behavior of Solutions to Cubic NLS Systems

Satoshi Masaki, Jun-ichi Segata, Kota Uriya · 2025

Read Non-Polynomial Conserved Quantities for ODE Systems and its Application to the Long-Time Behavior of Solutions to Cubic NLS Systems by Satoshi Masaki, Jun-ichi Segata, Kota Uriya (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-2501-13398.txt.

Introduction

Satoshi Masaki, Jun-ichi Segata, and Kota Uriya study the long-time behavior of waves governed by cubic nonlinear Schrodinger systems in a 2025 paper. These equations describe wave packets that interact as they travel, in optics and other fields. For small starting waves, the question is whether solutions stay bounded and what shape they settle into over long times.

The team finds a new class of systems where both answers come out well, even though the system lacks a usable conserved quantity of the usual kind. Their lever is a set of conserved quantities for a simpler reduced system derived from the wave system. Unlike an earlier study's polynomial versions, these quantities need not be polynomials at all. The authors use them to pin down the fading and form of small solutions.

Catalog record

CREATORSatoshi Masaki, Jun-ichi Segata, Kota Uriya
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE78 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-13398

Browse the full library · About the Tower of Babel

Full text

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

Non-Polynomial Conserved Quantities for ODE Systems and its Application to the Long-Time Behavior of Solutions to Cubic NLS Systems Satoshi Masaki, Jun-ichi Segata, Kota Uriya Abstract . In this paper, we investigate the asymptotic behavior of small solutions to the initial value problem for a system of cubic nonlinear Schrödinger equations (NLS) in one spatial dimension. We identify a new class of NLS systems for which the global boundedness and asymptotics of small solutions can be established, even in the absence of any effective conserved quantity. The key to this analysis lies in utilizing conserved quantities for the reduced ordinary differential equation (ODE) systems derived from the original NLS systems. In a previous study, the first author investigated conserved quantities expressed as quartic polynomials. In contrast, the conserved quantities considered in the present paper are of a different type and are not necessarily polynomial. 2010 Mathematics Subject Classification Primary 35Q55, Secondary 35A22, 35B40 1. Introduction In this paper, we consider the large time behavior of solutions to the initial value problem of the system of cubic nonlinear Schrödinger equations in one space dimension: (1.1) \left\{\begin{aligned} &i\partial_{t}u_{1}+\frac{1}{2}\partial_{x}^{2}u_{1}=F_{1}(u_{1},u_{2}),\quad t\in\mathbb{R},\ x\in\mathbb{R},\\ &i\partial_{t}u_{2}+\frac{1}{2}\partial_{x}^{2}u_{2}=F_{2}(u_{1},u_{2}),\quad t\in\mathbb{R},\ x\in\mathbb{R},\\ &u_{1}(0,x)=u_{1,0}(x),\quad u_{2}(0,x)=u_{2,0}(x),\qquad x\in\mathbb{R},\end{aligned}\right. where u_{j}:\mathbb{R}\times\mathbb{R}\to\mathbb{C} ( j=1,2 ) are unknown functions and u_{j,0}:\mathbb{R}\to\mathbb{C} ( j=1,2 ) are given functions. The nonlinear terms F_{1} and F_{2} are given by \displaystyle F_{1}(u_{1},u_{2}) \displaystyle=c_{1}|u_{1}|^{2}u_{1}+c_{2}|u_{1}|^{2}u_{2}+c_{3}u_{1}^{2}\overline{u_{2}}+c_{4}u_{1}|u_{2}|^{2}+c_{5}\overline{u_{1}}u_{2}^{2}+c_{6}|u_{2}|^{2}u_{2}, \displaystyle F_{2}(u_{1},u_{2}) \displaystyle=c_{7}|u_{1}|^{2}u_{1}+c_{8}|u_{1}|^{2}u_{2}+c_{9}u_{1}^{2}\overline{u_{2}}+c_{10}u_{1}|u_{2}|^{2}+c_{11}\overline{u_{1}}u_{2}^{2}+c_{12}|u_{2}|^{2}u_{2}, and c_{j}\ (j=1,\cdots,12) are real constants. Our aim is to show the global existence of the solution to ( 1.1 ) for small initial data and to derive their asymptotic behavior under suitable conditions on the parameters of the nonlinearities. It is well-known that the cubic nonlinearities are critical in determining the asymptotic behavior of solutions in one dimension, making this an important and intriguing problem. We provide two new types of systems of the form ( 1.1 ). A typical examples in our scope are (1.2) \left\{\begin{aligned} &i\partial_{t}u_{1}+\frac{1}{2}\partial_{x}^{2}u_{1}=(\zeta+1)|u_{1}|^{2}u_{2}+\zeta|u_{2}|^{2}u_{2},\quad t\in\mathbb{R},\ x\in\mathbb{R},\\ &i\partial_{t}u_{2}+\frac{1}{2}\partial_{x}^{2}u_{2}=2\zeta|u_{1}|^{2}u_{1}+2\zeta u_{1}|u_{2}|^{2}+\overline{u_{1}}u_{2}^{2},\quad t\in\mathbb{R},\ x\in\mathbb{R}\end{aligned}\right. with \zeta\in\mathbb{R}\setminus\{0\} and (1.3) \left\{\begin{aligned} &i\partial_{t}u_{1}+\frac{1}{2}\partial_{x}^{2}u_{1}=6|u_{1}|^{2}u_{1}+3|u_{1}|^{2}u_{2}+u_{1}^{2}\overline{u_{2}},\quad t\in\mathbb{R},\ x\in\mathbb{R},\\ &i\partial_{t}u_{2}+\frac{1}{2}\partial_{x}^{2}u_{2}=-u_{1}|u_{2}|^{2}-4|u_{2}|^{2}u_{2},\quad t\in\mathbb{R},\ x\in\mathbb{R}.\end{aligned}\right. Considering the single gauge invariant cubic nonlinear Schrödinger equation: (1.4) i\partial_{t}u+\frac{1}{2}\partial_{x}^{2}u=\lambda|u|^{2}u,\quad t\in\mathbb{R},\ x\in\mathbb{R} with \lambda\in\mathbb{C} , two typical asymptotic behavior of the solution are known, depending on whether \lambda\in\mathbb{C} is a real number or a pure imaginary number. When \lambda is a real number, Ozawa [ Oz ] showed the modified scattering of the small solution to ( 1.4 ), namely, for a suitable function u_{+} depending only on the spatial variable, (1.5) u(t,x)=(it)^{-\frac{1}{2}}e^{\frac{i|x|^{2}}{2t}}\widehat{u_{+}}\left(\frac{x}{t}\right)\exp(-i\lambda|\widehat{u_{+}}(\tfrac{x}{t})|\log t)+o(t^{-\frac{1}{2}}) in L^{\infty} as t\to\infty (see also [ GiOz , HaNa , HaNa2 , IfTa , KaPu , LS ] ). When \lambda is a pure imaginary number, Shimomura [ Shi ] showed that the equation has the dissipative property in one time direction (See also [ KiSh , HaNaSu ] ). Kita [ 11 ] showed a counter part of the above result, namely small data blow-up occurs in converse time direction. In studying the general system of cubic nonlinear Schödinger equations ( 1.1 ), we encounter the various situations depending on the coefficients of the nonlinearity [ KaSa , Kim , LiNiSaSu1 , LiNiSaSu2 , LiNiSaSu5 , LiSu , LiSu2 , MuPu , Ur ] . In the our previous studies [ MaSeUr1 , MaSeUr2 ] , we provided the classification of the system ( 1.1 ). First, we review it in detail. In [ MaSeUr1 ] , we showed that the system ( 1.1 ) can be identified with a pair of a matrix \mathscr{A}\in M_{3}(\mathbb{R})\simeq\mathbb{R}^{9} and a vector \mathscr{V}\in\mathbb{R}^{3} as follows: Given (c_{1},\cdots,c_{12})\in\mathbb{R}^{12} , we define (1.6) \mathscr{A}:=\begin{bmatrix}c_{2}-c_{3}&-c_{1}+c_{8}-c_{9}&-c_{7}\\ c_{5}&-c_{3}+c_{11}&-c_{9}\\ c_{6}&-c_{4}+c_{5}+c_{12}&-c_{10}+c_{11}\end{bmatrix} and (1.7) \mathscr{V}:=\begin{bmatrix}c_{8}-2c_{9}\\ \frac{1}{2}(-c_{2}+2c_{3}-c_{10}+2c_{11})\\ c_{4}-2c_{5}\end{bmatrix}. Conversely, for a given pair (\mathscr{A}=(a_{ij})_{1\leqslant i,j\leqslant 3},(q_{k})_{1\leqslant k\leqslant 3})\in M_{3}(\mathbb{R})\times\mathbb{R}^{3} , the system of the form ( 1.1 ) is given with the following nonlinearity (see [ MaSeUr1 ] ); (1.8) \left\{\begin{aligned} F_{1}(u_{1},u_{2})={}&-(a_{12}+a_{23})|u_{1}|^{2}u_{1}+a_{11}(2|u_{1}|^{2}u_{2}+u_{1}^{2}\overline{u_{2}})+a_{21}(2u_{1}|u_{2}|^{2}+\overline{u_{1}}u_{2}^{2})\\ &+a_{31}|u_{2}|^{2}u_{2}-(\tr\mathscr{A})\operatorname{Re}(\overline{u_{1}}u_{2})u_{1}+\mathcal{V}(u_{1},u_{2})u_{1},\\ F_{2}(u_{1},u_{2})={}&-a_{13}|u_{1}|^{2}u_{1}-a_{23}(2|u_{1}|^{2}u_{2}+u_{1}^{2}\overline{u_{2}})-a_{33}(2u_{1}|u_{2}|^{2}+\overline{u_{1}}u_{2}^{2})\\ &+(a_{21}+a_{32})|u_{2}|^{2}u_{2}+(\tr\mathscr{A})\operatorname{Re}(\overline{u_{1}}u_{2})u_{2}+\mathcal{V}(u_{1},u_{2})u_{2},\end{aligned}\right. where \mathcal{V}(u_{1},u_{2})=q_{1}|u_{1}|^{2}+2q_{2}\operatorname{Re}(\overline{u_{1}}u_{2})+q_{3}|u_{2}|^{2} is a real-valued quadratic potential. The matrix \mathscr{A} naturally appears in the analysis of the following system of ordinary differential equations, which is obtained by removing the term \frac{1}{2}\partial_{x}^{2}u_{j} from the system ( 1.1 ): (1.9) \left\{\begin{aligned} i\partial_{t}\varphi_{1}=F_{1}(\varphi_{1},\varphi_{2}),\\ i\partial_{t}\varphi_{2}=F_{2}(\varphi_{1},\varphi_{2}).\end{aligned}\right. The system ( 1.9 ) is called the reduced ODE system . We introduce the four quadratic quantities (1.10) \rho_{1}=|\varphi_{1}|^{2},\quad\rho_{2}=|\varphi_{2}|^{2},\quad\mathcal{R}=2\operatorname{Re}(\overline{\varphi_{1}}\varphi_{2}),\quad\mathcal{I}=2\operatorname{Im}(\overline{\varphi_{1}}\varphi_{2}) for a solution (\varphi_{1},\varphi_{2}) to ( 1.9 ). Further, for a vector \vec{a}={}^{t}\!(a_{1},a_{2},a_{3})\in\mathbb{R}^{3} and a solution (\varphi_{1},\varphi_{2}) to ( 1.9 ), we define a quadratic form (1.11) \mathcal{Q}(\vec{a}):=\begin{bmatrix}\rho_{1}&\mathcal{R}&\rho_{2}\end{bmatrix}\vec{a}=\begin{bmatrix}\overline{\varphi_{1}}&\overline{\varphi_{2}}\end{bmatrix}\begin{bmatrix}a_{1}&a_{2}\\ a_{2}&a_{3}\end{bmatrix}\begin{bmatrix}\varphi_{1}\\ \varphi_{2}\end{bmatrix}. Following proposition provides the time evolution of the quadratic form \mathcal{Q}(\vec{a}) along the flow given by ( 1.9 ). Proposition 1.1 ( [ MaSeUr2 ] ) . For any \vec{a}\in\mathbb{R}^{3} , it holds that (1.12) \frac{d}{dt}\mathcal{Q}(\vec{a})=2\mathcal{I}\mathcal{Q}(\mathscr{A}\vec{a}). In particular, if \vec{a}\in\ker\mathscr{A} then \mathcal{Q}(\vec{a}) is independent of time. We introduce the following sets: (1.13) \displaystyle\mathcal{P}_{+}:=\{{}^{t}\!(a,b,c)\in\mathbb{R}^{3}\ |\ ac>b^{2}\}\subset\mathbb{R}^{3}, \displaystyle\mathcal{P}_{0}:=\{{}^{t}\!(a,b,c)\in\mathbb{R}^{3}\ |\ ac=b^{2}\}\subset\mathbb{R}^{3}, \displaystyle\mathcal{P}_{-}:=\{{}^{t}\!(a,b,c)\in\mathbb{R}^{3}\ |\ ac<b^{2}\}\subset\mathbb{R}^{3}. Note that if \vec{a}\in\mathcal{P}_{+} then the quadratic form \mathcal{Q}(\vec{a}) is sign-definite, i.e., |\mathcal{Q}(\vec{a})|\sim_{\vec{a}}|\varphi_{1}|^{2}+|\varphi_{2}|^{2} . From Proposition 1.1 , we see that \rank\mathscr{A} indicates the complexity of the system since the number of the conserved quadratic quantities of the reduced ODE system ( 1.9 ) is 3-\rank\mathscr{A} . In [ MaSeUr2 , Ma ] , the system ( 1.1 ) satisfying \rank\mathscr{A}\leqslant 2 are classified. Roughly speaking, there are the following seven types; (1) \rank\mathscr{A}=0 ; (2) \rank\mathscr{A}=1 and \ker\mathscr{A}\cap\mathcal{P}_{+}\neq\emptyset ; (3) \rank\mathscr{A}=1 , \ker\mathscr{A}\cap\mathcal{P}_{+}=\emptyset , and \ker\mathscr{A}\cap\mathcal{P}_{0}\neq\{0\} ; (4) \rank\mathscr{A}=1 , \ker\mathscr{A}\cap\mathcal{P}_{+}=\emptyset , and \ker\mathscr{A}\cap\mathcal{P}_{0}=\{0\} ; (5) \rank\mathscr{A}=2 and \ker\mathscr{A}\cap\mathcal{P}_{+}\neq\emptyset ; (6) \rank\mathscr{A}=2 and \ker\mathscr{A}\cap\mathcal{P}_{0}\neq\{0\} ; (7) \rank\mathscr{A}=2 and \ker\mathscr{A}\cap\mathcal{P}_{-}\neq\emptyset . The case (1) includes the Manakov system [ 20 ] which is an integrable system arising in the nonlinear optics. In the cases (1), (2) and (5), a structural condition called the weak null gauge condition is satisfied. As for ( 1.1 ), this condition is expressed as (1.14) \ker\mathscr{A}\cap\mathcal{P}_{+}\neq\emptyset (see [ Ma2 ] ). If this condition is satisfied, then one can find a conserved quadratic form which is equivalent to |\varphi_{1}(t)|^{2}+|\varphi_{2}(t)|^{2} in view of Proposition 1.1 . Hence, solutions to the reduced ODE system ( 1.9 ) are global and bounded. Kim [ Kim ] studied the global existence and time decay of solution to ( 1.1 ) under the weak null gauge condition (see also Li-Sunagawa [ LiSu , LiSu2 ] for ( 1.1 ) with derivative cubic nonlinearity). Furthermore, Katayama-Sakoda [ KaSa ] proved that, under the weak null gauge condition, the asymptotic behavior of the solution of the system ( 1.1 ) is given by (1.15) u_{j}(t,x)=(it)^{-\frac{1}{2}}e^{\frac{i|x|^{2}}{2t}}\varphi_{j}\left(\log t,\frac{x}{t}\right)+O(t^{-\delta}), for \delta>1/2 , where \varphi_{1} and \varphi_{2} are the solution to ( 1.9 ) (see also [ HLN ] for the two dimensional case). We note that they considered not only ( 1.1 ) but also more general N -coupled systems of cubic nonlinear Schrödinger equations, including derivatives in the nonlinearity. In our previous studies [ MaSeUr2 , KiMaSeUr ] , we showed a similar asymptotic results for several examples belonging to the case (3) or (6). These examples has the special structure; one equation of the system is the single cubic equation ( 1.4 ) and the other equation is linear with time dependent potential given by the first component. Our main motivation here is to study the case \rank\mathscr{A}=3 . We remark that the condition ( 1.14 ) holds only in the case \rank\mathscr{A}\leqslant 2 because \ker\mathscr{A}=\{0\} if \rank\mathscr{A}=3 . Recently,

Read the complete text