The Sasa&ndash;Satsuma equation with high-order discrete spectra in space-time solitonic regions: soliton resolution via the mixed $\bar{\partial }$-Riemann&ndash;Hilbert problem

Minghe Zhang; Zhenya Yan

doi:10.1088/1572-9494/ad361b

PDF(652 KB)

Communications in Theoretical Physics ›› 2024, Vol. 76 ›› Issue (6) : 65002. DOI: 10.1088/1572-9494/ad361b

The Sasa–Satsuma equation with high-order discrete spectra in space-time solitonic regions: soliton resolution via the mixed $\bar{\partial}$ -Riemann–Hilbert problem

Author information +

History +

Abstract

In this paper, we investigate the Cauchy problem of the Sasa–Satsuma (SS) equation with initial data belonging to the Schwartz space. The SS equation is one of the integrable higher-order extensions of the nonlinear Schrödinger equation and admits a 3 × 3 Lax representation. With the aid of the $\bar{\partial}$ -nonlinear steepest descent method of the mixed $\bar{\partial}$ -Riemann–Hilbert problem, we give the soliton resolution and long-time asymptotics for the Cauchy problem of the SS equation with the existence of second-order discrete spectra in the space-time solitonic regions.

Key words

Sasa–Satsuma equation / inverse scattering / $\bar{\partial}$ -Riemann–Hilbert problem / $\bar{\partial}$ steepest descent method / soliton resolution

Cite this article

EndNote

Ris (Procite)

Bibtex

Download Citations

Minghe Zhang, Zhenya Yan. The Sasa–Satsuma equation with high-order discrete spectra in space-time solitonic regions: soliton resolution via the mixed

\bar{\partial}

-Riemann–Hilbert problem[J]. Communications in Theoretical Physics, 2024, 76(6): 65002 https://doi.org/10.1088/1572-9494/ad361b

1. Introduction

In 1993, based on the scheme of the Riemann–Hilbert (RH) problem, Deift and Zhou discussed the long-time asymptotic behaviors of the solutions of the mKdV equation by combining classical Fourier analysis and the steepest descent method [1]. Based on this method, the long-time asymptotic behaviors of many integrable equations were explored, such as the sine-Gordon equation [2], the Korteweg-de Vries equation [3], the Camassa–Holm equation [4], the short pulse equation [5, 6], the Fokas–Lenells equation [7], the extended mKdV equation [8, 9], and so on.

As a development of the Deift–Zhou nonlinear steepest descent method, a powerful tool called the

\bar{\partial}

-steepest descent method was first proposed by Mclaughlin and miller to analyze the asymptotic behaviors of orthogonal polynomials [10, 11]. Later, this method was successfully used to analyze the long-time behaviors of solutions to integrable nonlinear wave equations, such as the focusing NLS equation [12, 13], the defocusing NLS equation [14, 15], the derivative NLS equation [16], the mKdV equation [17, 18], the fifth-order mKdV equation [19], the complex short pulse equation [20], the modified Camassa–Holm equation [21], the Novikov equation [22], etc.

As a new-type integrable high-order equation of the nonlinear Schrödinger equation, the Sasa–Satsuma (SS) equation was presented [23]

\begin{array}{rcl} q_{t} + q_{x x x} + 3 {| q |}^{2} q_{x} + 3 {({| q |}^{2} q)}_{x} = 0, (x, t) \in R \times R^{+}, \end{array}

(1.1)

which admits a 3 × 3 Lax pair

\begin{array}{rcl} Φ_{x} + i k σ Φ = U (x, t; k) Φ, Φ_{t} + 4 i k^{3} σ = W (x, t; k) Φ, \end{array}

(1.2)

where Φ = Φ(x, t; k) is a matrix function of x, t and iso-spectral parameter

k \in C

\begin{array}{rcl} \begin{array}{rcl} U (x, t; k) & = & (\begin{array}{cc} 0_{2 \times 2} & q (x, t) \\ - q^{†} (x, t) & 0_{1 \times 2} \end{array}), \\ q (x, t) & = & (\begin{array}{c} q (x, t) \\ q^{*} (x, t) \end{array}), \\ σ & = & (\begin{array}{cc} I_{2 \times 2} & 0_{2 \times 1} \\ 0_{1 \times 2} & - 1 \end{array}), \\ W (x, t; k) & = & 4 k^{2} U + 2 i k σ (U_{x} - U^{2}) \\ + 2 U^{3} - U_{x x} + [U_{x}, U], \end{array} \end{array}

(1.3)

with ‘*' and ‘†' denoting the complex conjugation and Hermite transformation, respectively. In fact, the SS equation (1.1) can also be regarded as the special reduction (r = q^*) of the two-component integrable complex modified KdV equations [24, 25]

\begin{array}{rcl} \begin{array}{l} q_{t} + q_{x x x} + 6 | q |^{2} q_{x} + 3 {(q r)}_{x} r^{*} = 0, \\ r_{t} + r_{x x x} + 6 | r |^{2} r_{x} + 3 {(q r)}_{x} q^{*} = 0. \end{array} \end{array}

(1.4)

The SS equation admits many other integrable properties, such as N-soliton solutions, infinite conservation laws, nonlocal symmetries, Painlevé property, dark soliton solutions, and rogue wave solutions [26–30]. Recently, Liu et al studied the long-time asymptotic behaviors of the SS equation via the Deift–Zhou nonlinear steepest descent method [31]. Recently, Xun and Fan used this method to study the long-time and Painlevé-type asymptotics of the SS equation under the assumption of scattering data admitting only finitely simple zeros [32].

Based on the above-mentioned situations, in this paper, we focus on the long-time asymptotic behaviors of solutions for the Cauchy problem of the integrable SS equation (1.1) with the initial data:

\begin{array}{rcl} q (x, 0) = q_{0} (x) \in S (R), \end{array}

(1.5)

under the assumption of the scattering data possessing finitely double zeros, where

S (R)

denotes the Schwartz space. We then obtain the long-time behaviors of the potential q(x, t).

The rest of this paper is organized as follows. In section 2, we review the direct and inverse scattering transforms about the 3 × 3 Lax pair of equation (1.1) and deduce the analytic region about the Jost functions. Furthermore, we set up the original RH problem. Based on the RH problem, in section 3, using the ideas from [13, 32], we give a series of the transformation of the RH problem to make it a model RH problem whose solution is a parabolic function. In section 4, through the transformations of the RH problem, the potential of RHP1 can be reconstructed by three parts. One is the double-pole soliton solutions by solving the RHP in the reflectionless case, and the other terms are provided by the error function E(r) and the pure

\bar{\partial}

-problem.

2. The direct scattering problem

2.1. Jost solutions of the Lax pair and scattering data

Based on the boundary-value condition

{l i m}_{| x | \to \infty} q_{0} (x) = 0

, the eigenfunction of the Lax pair (1.2) has the following asymptotic form

\begin{array}{rcl} Φ (k, x, t) \sim e^{- i (k x + 4 k^{3} t) σ}, | x | \to \infty . \end{array}

(2.1)

To change the large-space asymptotics of the eigenfunction of the Lax pair (1.2) into a unit matrix, let

\begin{array}{rcl} Ψ (k, x, t) = Φ (k, x, t) e^{i (k x + 4 k^{3} t) σ} . \end{array}

Then

Ψ

(k, x, t) satisfies the following modified Lax pair

\begin{array}{rcl} Ψ_{x} + i k [σ, Ψ] = U Ψ, Ψ_{t} + 4 i k^{3} [σ, Ψ] = W Ψ, \end{array}

(2.2)

which can be written as a fully differential form

\begin{array}{rcl} d (e^{i (k x + 4 k^{3} t) \hat{σ}} Ψ) = e^{i (k x + 4 k^{3} t) \hat{σ}} (U Ψ d x + W Ψ d t), \end{array}

(2.3)

from which the Jost solutions

Ψ

₊(k, x, t) and

Ψ

₋(k, x, t) can be rewritten as follows:

\begin{array}{rcl} Ψ_{\pm} (k, x, t) = I - \int_{x}^{\pm \infty} e^{i k (ξ - x) \hat{σ}} U (ξ, t) Ψ_{\pm} (k, y, t) d ξ . \end{array}

(2.4)

Let

Ψ_{\pm} = (Ψ_{\pm 1} (k, x, t), Ψ_{\pm 2} (k, x, t))

, where

Ψ

_±1(k, x, t) and

Ψ

_±2(k, x, t) represent their first two columns and third column, respectively. It follows from equation (2.4) that

Ψ

₋₁,

Ψ

₊₂ are analytic in

C_{+}

, and

Ψ

₊₁,

Ψ

₋₂ are analytic in

C_{-}

. Moreover,

\begin{array}{rcl} \begin{array}{l} (Ψ_{\mp 1} (k, x, t), Ψ_{\pm 2} (k, x, t)) = I + O (k^{- 1}), \\ k \in C_{\pm} \to \infty . \end{array} \end{array}

(2.5)

By using Abel's lemma and

t r (U) = t r (W) = 0

, one knows that

det Ψ_{\pm} (k, x, t)

are independent of variable x and

det Ψ_{\pm} = 1

. Furthermore,

Ψ_{\pm} e^{- i (k x + 4 k^{3} t) σ}

are linearly dependent to lead to

\begin{array}{rcl} Ψ_{-} (k) e^{- i (k x + 4 k^{3} t) σ} = Ψ_{+} (k) e^{- i (k x + 4 k^{3} t) σ} S (k), \end{array}

(2.6)

where S(k) is a 3 × 3 scattering matrix. Moreover, together with

det (S) = 1

, one can know that

Ψ

_± and S(k) admit the symmetries:

\begin{array}{rcl} \begin{array}{l} Ψ_{\pm}^{†} (k^{*}; x, t) = Ψ^{- 1} (k; x, t), \\ Ψ_{\pm} (k; x, t) = ϱ Ψ_{\pm}^{*} (- k^{*}, x, t) ϱ, \end{array} \end{array}

(2.7)

\begin{array}{rcl} \begin{array}{l} S^{†} (k^{*}) = S^{- 1} (k), \\ S (k) = ϱ S^{*} (- k^{*}) ϱ, \end{array} \end{array}

(2.8)

based on the two symmetries of U

\begin{array}{rcl} \begin{array}{rcl} U^{†} & = & - U, ϱ U ϱ = U^{*}, ϱ = ϱ^{- 1} = (\begin{array}{cc} σ_{1} & 0 \\ 0 & 1 \end{array}), \\ σ_{1} & = & (\begin{array}{cc} 0 & 1 \\ 1 & 0 \end{array}) . \end{array} \end{array}

(2.9)

One can further rewrite S(k) as

\begin{array}{rcl} \begin{array}{rcl} S (k) & = & (\begin{array}{cc} A (k) & - a d j [A^{†} (k^{*})] B^{†} (k^{*}) \\ B (k) & det [A^{†} (k^{*})] \end{array}) \\ = \underset{x \to \infty}{l i m} e^{i k x \hat{σ}} Ψ_{-} (k; x, 0), \end{array} \end{array}

(2.10)

where

A (k) = σ_{1} A^{*} (-- k^{*}) σ_{1} = {(a_{i j} (k))}_{2 \times 2}

, adj[A^†(k^*)] denotes the adjoint matrix of A^†(k^*), and B(k) = B^*(–k^*)σ₁ = (B₁(k), B₂(k)), and

\begin{array}{rcl} \begin{array}{rcl} A (k) & = & I + \int_{R} q (x, 0) Ψ_{- 12} (k; x, 0) d x, \\ B (k) & = & - \int_{R} q^{†} (x, 0) Ψ_{- 11} (k; x, 0) e^{- 2 i k x} d x, \end{array} \end{array}

(2.11)

which imply that A(k) is analytic in

C_{+}

by virtue of the analyticity of

Ψ

₋₁₂(k; x, 0).

For the convenience of the following analysis, an assumption of scattering data is that the functions A(k) and

det A (k)

have no zeros on

R

and A(k) has finite double zeros in

C ∖ R

γ (k) := B (k) A^{- 1} (k) \in H^{1, 1} (R)

2.2. The Riemann–Hilbert problem with higher-order poles

Let A(k) have 2N double zeros k₁, k₂,…,k_2N in

C_{+}

with

k_{N + j} = - k_{j}^{*}, j = 1, 2, \dots, N

since there is the symmetry A(k) = σ₁A^*(− k^*)σ₁, that is,

A (k_{j}) = A^{'} (k_{j}) = 0, A^{''} (k_{j}) \neq 0 (j = 1, 2, \dots, 2 N)

. To establish a RH problem, we define the following sectionally meromorphic matrix M(k; x, t) with the aid of the analyticity of Jost functions

\begin{array}{rcl} M (k; x, t) = {\begin{cases} (Ψ_{- 1} (k) A^{- 1} (k), Ψ_{+ 2} (k)), & k \in C^{+}, \\ (Ψ_{+ 1} (k), \frac{Ψ_{- 2} (k)}{det A^{†} (k^{*})}), & k \in C^{-}, \end{cases} \end{array}

(2.12)

such that M(k; x, t) has 2N double poles K = {k_j, j = 1,…,2N} in

C^{+}

and 2N double poles

\bar{K} = {k_{j}^{*}, j = 1, \dots, 2 N}

C^{-}

. According to equations (2.6) and (2.12), one can find that the matrix-valued function M(k; x, t) satisfies the following RH problem:

RHP-1. Find a matrix-valued function solution M(k; x, t) satisfying the following conditions:

•

•Analyticity : M(k; x, t) is a meromorphic function in $C ∖ R$ and has double poles at k_j ∈ K and $k_{j}^{*} \in \bar{K};$

•

•Jump relation: M(k; x, t) has continuous boundary values M_±(k; x, t) on $R$ , and

\begin{array}{rcl} M_{+} (k) = M_{-} (k) V (k; x, t), k \in R, \end{array}

(2.13)

where the jump matrix is

\begin{array}{rcl} \begin{array}{rcl} V (k; x, t) & = & (\begin{array}{cc} I_{2 \times 2} + γ^{†} (k^{*}) γ (k) & γ^{†} (k^{*}) e^{- 2 i t θ (k; x, t)} \\ γ (k) e^{2 i t θ (k; x, t)} & 1 \end{array}), \\ θ (k; x, t) & = & k (\frac{x}{t} + 4 k^{2}) . \end{array} \end{array}

(2.14)

•

•Asymptotics:

\begin{array}{rcl} M (k; x, t) = I + O (\frac{1}{k}), k \to \infty . \end{array}

(2.15)

Therefore, M(k; x, t) has double poles at each point in

K \cup \bar{K}

with:

\begin{array}{rcl} \underset{k = k_{j}}{P_{- 2}} M (k; x, t) = \underset{k \to k_{j}}{l i m} M (k; x, t) (\begin{array}{cc} 0 & 0 \\ A_{j} e^{2 i t θ (k)} & 0 \end{array}), \end{array}

(2.16)

\begin{array}{rcl} \begin{array}{l} {R e s}_{k = k_{j}} M (k; x, t) = \underset{k \to k_{j}}{l i m} M (k; x, t) (\begin{array}{cc} 0 & 0 \\ B_{j} e^{2 i t θ (k)} & 0 \end{array}) \\ + M^{'} (k; x, t) (\begin{array}{cc} 0 & 0 \\ A_{j} e^{2 i t θ (k)} & 0 \end{array}), \end{array} \end{array}

(2.17)

\begin{array}{rcl} \underset{k = k_{j}^{*}}{P_{- 2}} M (k; x, t) = \underset{k \to k_{j}^{*}}{l i m} M (k; x, t) (\begin{array}{cc} 0 & - A_{j}^{†} e^{- 2 i t θ (k)} \\ 0 & 0 \end{array}), \end{array}

(2.18)

\begin{array}{rcl} \begin{array}{l} {R e s}_{k = k_{j}^{*}} M (k; x, t) = \underset{k \to k_{j}^{*}}{l i m} M (k; x, t) (\begin{array}{cc} 0 & - B_{j}^{†} e^{- 2 i t θ (k)} \\ 0 & 0 \end{array}) \\ + M^{'} (k; x, t) (\begin{array}{cc} 0 & - A_{j}^{†} e^{- 2 i t θ (k)} \\ 0 & 0 \end{array}), \end{array} \end{array}

(2.19)

where

\begin{array}{rcl} \begin{array}{rcl} A_{j} & = & \frac{2 B (k_{j}) a d j [A (k_{j})]}{\ddot{det} [A (k_{j})]}, \\ B_{j} & = & (2 i t θ^{'} (k_{j}) + \frac{2}{3} \frac{\overset{⃛}{det} [A (k_{j})]}{\ddot{det} [A (k_{j})]}) A_{j} + A_{j}^{'} . \end{array} \end{array}

(2.20)

We now give the reconstruction formula for the solution of (1.1). Let

Ψ

have the following Laurent expansion

\begin{array}{rcl} Ψ = Ψ^{(0)} + \frac{Ψ^{(1)}}{k} + \frac{Ψ^{(2)}}{k^{2}} + \dots \end{array}

(2.21)

Then, by substituting (2.21) into (2.2) and comparing the k⁻² in t-part and k⁰ in the x part, one can obtain

\begin{array}{rcl} 4 i [σ, Ψ^{(1)}] = 4 U Ψ^{(0)}, Ψ_{x}^{(0)} + i [σ, Ψ^{(1)}] = U Ψ^{(0)} . \end{array}

(2.22)

From the above two equations, we have

\begin{array}{rcl} U = i [σ, Ψ^{(1)}], \end{array}

(2.23)

i.e.

\begin{array}{rcl} \begin{array}{rcl} q (x, t) & = & {(q (x, t), q^{*} (x, t))}^{T} = 2 i Ψ_{12}^{(1)} \\ = 2 i \underset{k \to \infty}{l i m} {(k M (k; x, t))}_{12}, \end{array} \end{array}

(2.24)

where q(x, t) solves the SS equation (1.1).

In the following, we mainly consider the solution of M(k; x, t) of RHP-1.

3. The mixed $\bar{\partial}$ -RH problem and its decomposition

3.1. Two factorizations of jump matrix V(k)

Since the jump matrix V(k; x, t) given by equation (2.14) admits the two different oscillatory terms for t > 0

\begin{array}{rcl} \begin{array}{rcl} O_{\pm} & = & e^{\pm 2 i t θ (k)} = e^{\pm 2 i t (4 k^{3} + \frac{x}{t} k)}, \\ θ (k) & = & \frac{k x}{t} + 4 k^{3} = 4 (k^{3} - 3 k_{0}^{2} k), \end{array} \end{array}

(3.1)

where θ(k) admits two phase points

k = \pm k_{0}, k_{0} = \sqrt{- \frac{x}{12 t}}

with xt < 0. To analyze their properties, one needs to consider the properties of

R e [i θ (k)]

of O_±

\begin{array}{rcl} R e [i θ (k)] = 4 I m k ({I m}^{2} k - 3 {R e}^{2} k + 3 k_{0}^{2}), \end{array}

(3.2)

whose signature table is given in figure 1.

Figure 1. The signature table of $R e i θ (k) = 4 I m k ({I m}^{2} k - 3 {R e}^{2} k + 3 k_{0}^{2})$ with ±k₀ being phase points.

Full size|PPT slide

To analyze the long-time asymptotics of RHP-1, we first divide all the poles into two parts:

\begin{array}{rcl} \begin{array}{rcl} Δ^{-} & = & {k | {R e}^{2} (k) - \frac{1}{3} {I m}^{2} (k) < k_{0}^{2}}, \\ Δ^{+} & = & {k | {R e}^{2} (k) - \frac{1}{3} {I m}^{2} (k) > k_{0}^{2}} . \end{array} \end{array}

(3.3)

We assume that there are no poles corresponding to the region Δ⁻ for simplicity.

The jump matrix V(k, x, t) has two different decompositions of upper and lower triangular matrices:

\begin{array}{rcl} \begin{array}{c} V = {\begin{array}{cc} (\begin{array}{cc} I & γ^{†} (k^{*}) e^{- 2 i t θ (k)} \\ 0 & 1 \end{array}) (\begin{array}{cc} I & 0 \\ γ (k) e^{2 i t θ (k)} & 1 \end{array}), & k \in R ∖ [- k_{0}, k_{0}], \\ (\begin{array}{cc} I & 0 \\ \frac{γ (k)}{1 + γ (k) γ^{†} (k^{*})} e^{2 i t θ (k)} & 1 \end{array}) (\begin{array}{cc} I + γ^{†} (k^{*}) γ (k) & 0 \\ 0 & \frac{1}{1 + γ (k) γ^{†} (k^{*})} \end{array}) (\begin{array}{cc} I & \frac{γ^{†} (k^{*})}{1 + γ (k) γ^{†} (k^{*})} e^{- 2 i t θ (k)} \\ 0 & 1 \end{array}), & k \in (- k_{0}, k_{0}) . \end{array} \end{array} \end{array}

(3.4)

To offset the influence of the diagonal matrix of the second decomposition, one needs to introduce the 2 × 2 matrix function δ(k) satisfying the following property:

Proposition 1.The matrix function $δ (k)$ and scalar function $det δ (k)$ satisfy the following properties:

•	• $δ (k)$ and $det (δ (k))$ are analytic, and $δ (k) δ^{†} (k^{}) = I$ , $det (δ (k)) det (δ^{} (k^{*})) = 1$ in $C ∖ [- k_{0}, k_{0}]$ .
•	•For $k \in (- k_{0}, k_{0})$ , $\begin{array}{rcl} \begin{array}{l} δ_{+} (k) = δ_{-} (k) (1 + γ^{†} (k) γ (k)), \\ det (δ_{+} (k)) = det (δ_{-} (k)) (1 + \| γ (k) \|^{2}); \end{array} \end{array}$ (3.5)
•	• $\begin{array}{rcl} \| δ_{+} (k) \|^{2} = {\begin{cases} \| γ (k) \|^{2} + 2, & k \in (- k_{0}, k_{0}), \\ 2, & o t h e r w i s e . \end{cases} \end{array}$ (3.6) $\begin{array}{rcl} \| δ_{-} (k) \|^{2} = {\begin{cases} 2 - \frac{\| γ (k) \|^{2}}{1 + \| γ (k) \|^{2}}, & k \in (- k_{0}, k_{0}), \\ 2, & o t h e r w i s e . \end{cases} \end{array}$ (3.7)
•	•As $\| k \| \to \infty$ with $\| \arg (k) \| ⩽ c < π$ , $\begin{array}{rcl} \begin{array}{l} δ (k) = I + O (k^{- 1}), det (δ (k)) = 1 \\ + \frac{i}{k} [\frac{1}{2 π} \int_{- k_{0}}^{k_{0}} l o g (\frac{1 + \| γ (ξ) \|^{2}}{1 + \| γ (k_{0}) \|^{2}}) d ξ - 2 ν k_{0}] \\ + O (k^{- 2}); \end{array} \end{array}$ (3.8)
•	• $\begin{array}{rcl} det (δ (k)) = {(\frac{k - k_{0}}{k + k_{0}})}^{i ν (k_{0})} e^{X (k)}, \end{array}$ (3.9) where $\begin{array}{rcl} \begin{array}{rcl} ν (k_{0}) & = & - \frac{1}{2 π} l o g (1 + \| γ (k_{0}) \|^{2}), \\ X (k) & = & \frac{1}{2 π i} \int_{- k_{0}}^{k_{0}} l o g (\frac{1 + \| γ (ξ) \|^{2}}{1 + \| γ (k_{0}) \|^{2}}) \frac{d ξ}{ξ - k} . \end{array} \end{array}$ (3.10)
•	•Along the ray $k = \pm k_{0} + R^{+} e^{i ϕ}$ with $\| ϕ \| ⩽ c < π$ , as $k \to \pm k_{0}$ , $\begin{array}{rcl} \| det (δ (k)) - {(\frac{k - k_{0}}{k + k_{0}})}^{i ν (k_{0})} e^{X (\pm k_{0})} \| ≲ \| k \mp k_{0} \|^{1 / 2} . \end{array}$ (3.11)

Proof.The proof of the above properties is similar to the proof of proposition 3.1 in [12].

In what follows, our aim is to find a transform of M(k; x, t) → M⁽¹⁾(k; x, t) such that the jump matrix of M⁽¹⁾(k; x, t) can be well decomposed. Let

\begin{array}{rcl} M^{(1)} (k; x, t) = M (k; x, t) T (k), \end{array}

(3.12)

where

\begin{array}{rcl} T (k) = (\begin{array}{cc} δ^{- 1} (k) & 0 \\ 0 & det [δ (k)] \end{array}) = (\begin{array}{cc} T_{1}^{- 1} (k) & 0 \\ 0 & T_{2} (k) \end{array}) \end{array}

(3.13)

with

δ (k), det [δ (k)]

are given by proposition 1, then the modified M⁽¹⁾(k; x, t) satisfies the following RHP-2:

RHP-2. Find a matrix-valued function M⁽¹⁾(k; x, t) satisfying

•

•Analyticity: M⁽¹⁾(k) is analytic in $C ∖ (R \cup K \cup \bar{K}) .$

•

•Jump condition: $M_{+}^{(1)} (k) = M_{-}^{(1)} (k) V^{(1)} (k), k \in R$ , where the jump matrix is

\begin{array}{rcl} V^{(1)} (k) = {\begin{array}{cc} (\begin{array}{cc} I & T_{1} T_{2} γ^{†} (k) e^{- 2 i t θ (k)} \\ 0 & 1 \end{array}) (\begin{array}{cc} I & 0 \\ {(T_{1} T_{2})}^{- 1} γ (k) e^{2 i t θ (k)} & 1 \end{array}), & k \in R ∖ [- k_{0}, k_{0}], \\ (\begin{array}{cc} I & 0 \\ \frac{{(T_{1 -} T_{2 -})}^{- 1} γ (k)}{1 + γ (k) γ^{†} (k)} e^{2 i t θ (k)} & 1 \end{array}) (\begin{array}{cc} I & \frac{T_{1 +} T_{2 +} γ^{†} (k)}{1 + γ (k) γ^{†} (k)} e^{- 2 i t θ (k)} \\ 0 & 1 \end{array}), & k \in (- k_{0}, k_{0}) . \end{array} \end{array}

(3.14)

•

•Asymptotics: $M^{(1)} (k) = I + O (k^{- 1}), a s k \to \infty$ .

Moreover, M⁽¹⁾(k; x, t) satisfies the following residue conditions at double poles k_j ∈ K and

k_{j}^{*} \in \bar{K}

\begin{array}{rcl} \underset{k = k_{j}}{P_{- 2}} M^{(1)} (k; x, t) = \underset{k \to k_{j}}{l i m} M^{(1)} (k) (\begin{array}{cc} 0 & 0 \\ A_{j} T_{1}^{- 1} T_{2}^{- 1} e^{2 i t θ (k)} & 0 \end{array}), \end{array}

(3.15)

\begin{array}{rcl} \begin{array}{l} {R e s}_{k = k_{j}} M^{(1)} (k; x, t) = \underset{k \to k_{j}}{l i m} M^{(1)} (k) \\ \times (\begin{array}{cc} 0 & 0 \\ (B_{j} T_{1}^{- 1} T_{2}^{- 1} + A_{j} T_{1}^{- 1} {T_{2}^{- 1}}^{'}) e^{2 i t θ (k)} & 0 \end{array}) \\ + {M^{(1)}}^{'} (k) (\begin{array}{cc} 0 & 0 \\ A_{j} T_{1}^{- 1} T_{2}^{- 1} e^{2 i t θ (k)} & 0 \end{array}), \end{array} \end{array}

(3.16)

\begin{array}{rcl} \underset{k = k_{j}^{*}}{P_{- 2}} M^{(1)} (k; x, t) = \underset{k \to k_{j}^{*}}{l i m} M^{(1)} (\begin{array}{cc} 0 & - A_{j}^{†} T_{1} T_{2} e^{2 i t θ} \\ 0 & 0 \end{array}), \end{array}

(3.17)

\begin{array}{rcl} \begin{array}{l} {R e s}_{k = k_{j}^{*}} M^{(1)} (k; x, t) = \underset{k \to k_{j}^{*}}{l i m} M^{(1)} (k) \\ \times (\begin{array}{cc} 0 & (- B_{j}^{†} T_{1} T_{2} - A_{j}^{†} T_{2} T_{1}^{'}) e^{- 2 i t θ (k)} \\ 0 & 0 \end{array}) \\ + {M^{(1)}}^{'} (k) (\begin{array}{cc} 0 & - A_{j}^{†} T_{1} T_{2} e^{- 2 i t θ (k)} \\ 0 & 0 \end{array}) . \end{array} \end{array}

(3.18)

To open the original jump curve

R

along the steepest descent lines arising from the phase points ±k₀, let these contours be

\begin{array}{rcl} \begin{array}{rcl} Σ_{1}^{\pm} & = & {k | k \mp k_{0} = R^{+} e^{i \frac{(2 \pm 1) π}{4}}}, \\ Σ_{3}^{\pm} & = & {k | k \mp k_{0} = d e^{i \frac{(2 \pm 1) π}{4}}, d \in (0, \sqrt{2} k_{0})}, \\ Σ_{2}^{\pm} & = & {k | k \mp k_{0} = R^{+} e^{i \frac{(- 2 \pm 1) π}{4}}}, \\ Σ_{4}^{\pm} & = & {k | k \mp k_{0} = d e^{i \frac{(- 2 \mp 1) π}{4}}, d \in (0, \sqrt{2} k_{0})} . \end{array} \end{array}

(3.19)

Then the complex plane

C

is divided into ten open domains, denoted by

Ω_{j}^{\pm}, j = 1, 2, 3, 4

and Ω₅, Ω₆ (see figure 2). In what follows, we will introduce the continuous functions related to the jump matrix V⁽¹⁾ in these regions.

Figure 2. Deformation of the jump countor from $R$ to Σ⁽²⁾.

Full size|PPT slide

Figure 3. Pole distribution. The red, green and yellow points generate the breather solutions. Moreover, the red points lie in the region $K^{+} (I)$ , the green points lie in the region $K (I)$ , and the yellow points on the line $R e i θ (k) = 0$ .

Full size|PPT slide

Figure 4. The jump contour for the jump matrix V⁽ⁱⁿ⁾.

Full size|PPT slide

Figure 5. The jump contour Σ^(E) for the error function E(k).

Full size|PPT slide

Proposition 2.Let $D = (- k_{0}, k_{0}), D_{-} = (- \infty, - k_{0}), D_{+} = (k_{0}, + \infty)$ . Then there exists the continuous functions $R_{j}^{\pm}$ : ${\bar{Ω}}_{j}^{\pm} \to C$ , $j = 1, 2, 3, 4$ such that

\begin{array}{rcl} R_{1}^{\pm} (k) = {\begin{cases} \cos (2 \arctan (k \mp k_{0})) g_{1}^{\pm} + [1 - \cos (2 \arctan (k \mp k_{0}))] f_{1}^{\pm}, & k \in {\bar{Ω}}_{1}^{\pm}, \\ g_{1}^{\pm} = - γ (k) T_{1}^{- 1} (k) T_{2}^{- 1} (k), & k \in D_{\pm}, \\ f_{1}^{\pm} = - γ (\pm k_{0}) T_{1}^{- 1} (k) e^{- X (\pm k_{0})} {(\frac{k - k_{0}}{k + k_{0}})}^{- i ν} (1 - X_{K} (k)), & k \in Σ_{1}^{\pm}, \end{cases} \end{array}

(3.20)

\begin{array}{rcl} R_{2}^{\pm} (k) = {\begin{cases} \cos (2 \arctan (k \mp k_{0})) g_{2}^{\pm} + [1 - \cos (2 \arctan (k \mp k_{0}))] f_{2}^{\pm}, & k \in {\bar{Ω}}_{2}^{\pm}, \\ g_{2}^{\pm} = T_{1} (k) T_{2} (k) γ^{†} (k), & k \in D_{\pm}, \\ f_{2}^{\pm} = T_{1} (k) e^{X (\pm k_{0})} {(\frac{k - k_{0}}{k + k_{0}})}^{i ν} γ^{†} (\pm k_{0}) (1 - X_{K} (k)), & k \in Σ_{2}^{\pm}, \end{cases} \end{array}

(3.21)

\begin{array}{rcl} R_{3}^{\pm} (k) = {\begin{cases} \cos (2 \arctan (k \mp k_{0})) g_{3} + [1 - \cos (2 \arctan (k \mp k_{0}))] f_{3}^{\pm}, & k \in {\bar{Ω}}_{3}^{\pm}, \\ g_{3} = - T_{1 +} (k) T_{2 +} (k) \frac{γ^{†} (k)}{1 + γ (k) γ^{†} (k)}, & k \in D, \\ f_{3}^{\pm} = - T_{1} (k) e^{X (\pm k_{0})} {(\frac{k - k_{0}}{k + k_{0}})}^{i ν} \frac{γ^{†} (\pm k_{0})}{1 + γ (\pm k_{0}) γ^{†} (\pm k_{0})} (1 - X_{K} (k)), & k \in Σ_{3}^{\pm}, \end{cases} \end{array}

(3.22)

\begin{array}{rcl} R_{4}^{\pm} (k) = {\begin{cases} \cos (2 \arctan (k \mp k_{0})) g_{4} + [1 - \cos (2 \arctan (k \mp k_{0}))] f_{4}^{\pm}, & k \in {\bar{Ω}}_{4}^{\pm}, \\ g_{4} = \frac{γ (k)}{1 + γ (k) γ^{†} (k)} T_{1 -}^{- 1} (k) T_{2 -}^{- 1} (k), & k \in D, \\ f_{4}^{\pm} = \frac{γ (\pm k_{0})}{1 + γ (\pm k_{0}) γ^{†} (\pm k_{0})} T_{1}^{- 1} (k) e^{- X (\pm k_{0})} {(\frac{k - k_{0}}{k + k_{0}})}^{- i ν} (1 - X_{K} (k)), & k \in Σ_{4}^{\pm}, \end{cases} \end{array}

(3.23)

and

R_{j}^{\pm} (k) (j = 1, 2, 3, 4)

have these estimates

\begin{array}{rcl} \begin{array}{l} | R_{j}^{\pm} (k) | ≲ 1 + ⟨ R e (k) ⟩^{- 1 / 2}, \\ | \bar{\partial} R_{j}^{\pm} (k) | ≲ | \bar{\partial} χ_{K} (k) | + | h_{j}^{'} (k) (R e (k)) | + | k \mp k_{0} |^{- 1 / 2}, \\ \bar{\partial} R_{j} \pm (k) = 0, f o r k \in Ω_{5} \cup Ω_{6} o r d i s t (k, K \cup \bar{K}) < ρ / 3, \end{array} \end{array}

(3.24)

where

\begin{array}{rcl} \begin{array}{rcl} h_{1} (k) & = & γ (k), h_{2} (k) = γ^{†} (k^{*}), \\ h_{3} (k) & = & \frac{h_{2} (k)}{1 + h_{1} (k) h_{2} (k)}, \\ h_{4} (k) & = & \frac{h_{1} (k)}{1 + h_{1} (k) h_{2} (k)} . \end{array} \end{array}

and

\begin{array}{rcl} \begin{array}{l} ρ = \frac{1}{2} min_{λ \neq μ \in K \cup \bar{K}} | λ - μ |, X_{K} (k) \\ = {\begin{array}{cc} 1, & d i s t (k, K \cup \bar{K}) < \frac{ρ}{3}, \\ 0, & d i s t (k, K \cup \bar{K}) > \frac{2 ρ}{3} . \end{array} \end{array} \end{array}

Proof.The proof is similar to [12, 32].

Let

\begin{array}{rcl} M^{(2)} (k; x, t) = M^{(1)} (k; x, t) R^{(2)} (k; x, t), \end{array}

(3.25)

where R⁽²⁾ is defined as

\begin{array}{rcl} R^{(2)} (k) = {\begin{array}{cc} (\begin{array}{cc} I & 0 \\ R_{j}^{\pm} (k) e^{2 i t θ (k)} & 1 \end{array}), & j = 1, 3, \\ (\begin{array}{cc} I & R_{j}^{\pm} (k) e^{- 2 i t θ (k)} \\ 0 & 1 \end{array}), & j = 2, 4, \\ I_{3 \times 3}, & o t h e r w i s e \end{array} \end{array}

(3.26)

with

R_{j}^{\pm} (k)

being defined by proposition 2 (notice that the transform causes the previous contour

R

to change into contour Σ⁽²⁾). Then M⁽²⁾(k; x, t) solves a mixed

\bar{\partial}

-RH problem:

Mixed

\bar{\partial}

-RHP. Find a matrix-valued function M⁽²⁾(k) = M⁽²⁾(k; x, t) solving

•

•Continuity: M⁽²⁾(k) is continuous in $C ∖ (Σ^{(2)} \cup K \cup \bar{K})$ .

•

•Jump condition: $M_{+}^{(2)} (k) = M_{-}^{(2)} (k) V^{(2)} (k), k \in Σ^{(2)}$ , where

\begin{array}{rcl} V^{(2)} (k) = {\begin{cases} (\begin{array}{cc} I & 0 \\ {(- 1)}^{j} R_{j}^{\pm} e^{2 i t θ} & 1 \end{array}), k \in Σ_{j}^{\pm}, j = 1, 4, \\ (\begin{array}{cc} I & {(- 1)}^{j} R_{j}^{\pm} e^{- 2 i t θ} \\ 0 & 1 \end{array}), k \in Σ_{j}^{\pm}, j = 2, 3, \\ (\begin{array}{cc} I & (R_{3}^{+} - R_{3}^{-}) e^{2 i t θ} \\ 0 & 1 \end{array}), k \in (i k_{0} \tan (\frac{π}{12}), i k_{0}), \\ (\begin{array}{cc} I & 0 \\ (R_{4}^{+} - R_{4}^{-}) e^{2 i t θ} & 1 \end{array}), k \in (- i k_{0} \tan (\frac{π}{12}), - i k_{0}), \\ I_{3 \times 3}, k \in (- i k_{0} \tan (\frac{π}{12}), i k_{0} \tan (\frac{π}{12})) . \end{cases} \end{array}

(3.27)

•

•Asymptotics: M⁽²⁾(k) → I, k → ∞ ;

Moreover, for any

k \in C ∖ (Σ^{(2)} ⋃ K ⋃ \bar{K})

, one finds that

\begin{array}{rcl} \bar{\partial} M^{(2)} (k) = M^{(1)} (k) \bar{\partial} R^{(2)} (k), \end{array}

(3.28)

where

\begin{array}{rcl} \bar{\partial} R^{(2)} (k) = {\begin{array}{cc} (\begin{array}{cc} 0 & 0 \\ \bar{\partial} R_{j}^{\pm} (k) e^{2 i t θ (k)} & 0 \end{array}), & j = 1, 3, \\ (\begin{array}{cc} 0 & \bar{\partial} R_{j}^{\pm} (k) e^{- 2 i t θ (k)} \\ 0 & 0 \end{array}), & j = 2, 4, \\ 0_{3 \times 3}, & o t h e r w i s e, \end{array} \end{array}

(3.29)

and, M⁽²⁾(k; x, t) has double poles at k_j and

k_{j}^{*}

with

\begin{array}{rcl} \underset{k = k_{j}}{P_{- 2}} M^{(2)} (k; x, t) = \underset{k \to k_{j}}{l i m} M^{(2)} (k) (\begin{array}{cc} 0 & 0 \\ A_{j} T_{1}^{- 1} T_{2}^{- 1} e^{2 i t θ (k)} & 0 \end{array}), \end{array}

(3.30)

\begin{array}{rcl} \begin{array}{l} {R e s}_{k = k_{j}} M^{(2)} (k; x, t) = \underset{k \to k_{j}}{l i m} M^{(2)} (k) \\ \times (\begin{array}{cc} 0 & 0 \\ (B_{j} T_{1}^{- 1} T_{2}^{- 1} + A_{j} T_{1}^{- 1} (T_{2}^{- 1})^{'}) e^{2 i t θ (k)} & 0 \end{array}) \\ + (M^{(2)})^{'} (k) (\begin{array}{cc} 0 & 0 \\ A_{j} T_{1}^{- 1} T_{2}^{- 1} e^{2 i t θ (k)} & 0 \end{array}), \end{array} \end{array}

(3.31)

\begin{array}{rcl} \underset{k = k_{j}^{*}}{P_{- 2}} M^{(2)} (k; x, t) = \underset{k \to k_{j}^{*}}{l i m} M^{(2)} (\begin{array}{cc} 0 & {\hat{A}}_{j} T_{1} T_{2} e^{2 i t θ} \\ 0 & 0 \end{array}), \end{array}

(3.32)

\begin{array}{rcl} \begin{array}{l} {R e s}_{k = k_{j}^{*}} M^{(2)} (k; x, t) = \underset{k \to k_{j}^{*}}{l i m} M^{(2)} (k) \\ \times (\begin{array}{cc} 0 & ({\hat{B}}_{j} T_{1} T_{2} + {\hat{A}}_{j} T_{2} (T_{1})^{'}) e^{- 2 i t θ (k)} \\ 0 & 0 \end{array}) \\ + (M^{(2)})^{'} (k) (\begin{array}{cc} 0 & {\hat{A}}_{j} T_{1} T_{2} e^{- 2 i t θ (k)} \\ 0 & 0 \end{array}) . \end{array} \end{array}

(3.33)

3.2. Analysis on a pure RH problem

Throughout this section, our aim is to decompose the above-mentioned mixed

\bar{\partial}

-RHP into a pure RHP with

\bar{\partial} R^{(2)} = 0

and a pure

\bar{\partial}

-problem with

\bar{\partial} R^{(2)} \neq 0

. The decomposition of M⁽²⁾(k; x, t) can be given as follows:

\begin{array}{rcl} M^{(2)} (k; x, t) = {\begin{cases} M_{R H P}^{(2)} (k; x, t), a s \bar{\partial} R^{(2)} = 0, \\ M^{(3)} (k; x, t) M_{R H P}^{(2)} (k; x, t), a s \bar{\partial} R^{(2)} \neq 0, \end{cases} \end{array}

(3.34)

where

M_{R H P}^{(2)} (k; x, t)

and M⁽³⁾(k; x, t) correspond to the pure RHP part and the pure

\bar{\partial}

part without jumps and poles of M⁽²⁾(k), respectively.

RHP-3. Find a matrix-valued function

M_{R H P}^{(2)} (k)

solving the following RHP

•	•Analyticity: $M_{R H P}^{(2)} (k)$ is analytic in $C ∖ (Σ^{(2)} \cup K \cup \bar{K})$ .
•	•Jump condition: $M_{R H P +}^{(2)} (k) = M_{R H P -}^{(2)} (k) V^{(2)} (k), k \in Σ^{(2)}$ , where V⁽²⁾(k) is given by equation (3.27).
•	•Asymptotics: $M_{R H P}^{(2)} (k) \to I$ , k → ∞ .

Proposition 3.The jump matrix $V^{(2)}$ has the following estimate:

\begin{array}{rcl} \begin{array}{l} ∥ V^{(2)} (k; x, t) - I ∥_{L^{\infty} (Σ^{(2)})} \\ = {\begin{cases} O (e^{- 6 k_{0} ρ^{2} t}), & k \in Σ_{j}^{\pm} ∖ \partial U_{\pm k_{0}}, j = 1, 2, \\ O (e^{- 8 k_{0}^{2} ρ t}), & k \in Σ_{j}^{\pm} ∖ \partial U_{\pm k_{0}}, j = 3, 4, \\ O ({(k_{0} | k \mp k_{0} |)}^{- 1} t^{- 1 / 2}), & k \in Σ^{(2)} ⋂ U_{\pm k_{0}}, \\ O (e^{- 14 k_{0}^{3} \tan^{3} (\frac{π}{12}) t}), & k \in [\pm i k_{0}, \pm i k_{0} \tan (\frac{π}{12})], \\ 0, & k \in [- i k_{0} \tan (\frac{π}{12}), i k_{0} \tan (\frac{π}{12})], \end{cases} \end{array} \end{array}

(3.35)

where

U_{\pm k_{0}} = {k | | k \pm k_{0} | < ρ / 2} .

3.2.1. Soliton solutions corresponding to discrete spectra

In order to analyze the leading term of the solution, we firstly consider RHP-1. RHP-1 reduces to the following RH problem:

RHP-4. A matrix-valued function M(k; x, t∣σ_d) with the scattering data

σ_{d} = {(k_{j}, A_{j}, B_{j})}_{k = 1}^{2 N}

and

K = {k_{j}}_{j = 1}^{2 N}

and satisfies the following condition:

•

•Analyticity: M(k; x, t∣σ_d) is analytical in $C ∖ (Σ^{(2)} \cup K \cup \bar{K}) .$

•

•Jump condition:

\begin{array}{rcl} M_{+} (k; x, t | σ_{d}) = M_{-} (k; x, t | σ_{d}) V (k), \end{array}

(3.36)

•

•Asymptotics: $M (k; x, t | σ_{d}) = I + O (k^{- 1})$ , k → ∞ .

Moreover, M(k; x, t∣σ_d) has double poles at k_j and

k_{j}^{*}

with

\begin{array}{rcl} \underset{k = k_{j}}{P_{- 2}} M (k; x, t | σ_{d}) = \underset{k \to k_{j}}{l i m} M (k; x, t | σ_{d}) (\begin{array}{cc} 0 & 0 \\ A_{j} e^{2 i t θ (k)} & 0 \end{array}), \end{array}

(3.37)

\begin{array}{rcl} \begin{array}{rcl} {R e s}_{k = k_{j}} M (k; x, t | σ_{d}) & = & \underset{k \to k_{j}}{l i m} M (k; x, t) (\begin{array}{cc} 0 & 0 \\ B_{j} e^{2 i t θ (k)} & 0 \end{array}) \\ + M^{'} (k; x, t | σ_{d}) (\begin{array}{cc} 0 & 0 \\ A_{j} e^{2 i t θ (k)} & 0 \end{array}) \end{array}, \end{array}

(3.38)

\begin{array}{rcl} \begin{array}{rcl} \underset{k = k_{j}^{*}}{P_{- 2}} M (k; x, t | σ_{d}) & = & \underset{k \to k_{j}^{*}}{l i m} M (k; x, t | σ_{d}) \\ \times (\begin{array}{cc} 0 & - A_{j}^{†} e^{- 2 i t θ (k)} \\ 0 & 0 \end{array}), \end{array} \end{array}

(3.39)

\begin{array}{rcl} \begin{array}{rcl} {R e s}_{k = k_{j}^{*}} M (k; x, t | σ_{d}) & = & \underset{k \to k_{j}^{*}}{l i m} M (k; x, t | σ_{d}) \\ \times (\begin{array}{cc} 0 & - B_{j}^{†} e^{- 2 i t θ (k)} \\ 0 & 0 \end{array}) \\ + M^{'} (k; x, t | σ_{d}) (\begin{array}{cc} 0 & - A_{j}^{†} e^{- 2 i t θ (k)} \\ 0 & 0 \end{array}) . \end{array} \end{array}

(3.40)

Proposition 4.Given scattering data $σ_{d} = (k_{j}, A_{j}, B_{j}))_{k = 1}^{2 N}$ and discrete spectra $K = {k_{j}}_{j = 1}^{2 N}$ , the RH problem has a unique solution

\begin{array}{rcl} q_{s o l} (x, t | σ_{d}) = {(q (x, t), q^{*} (x, t))}^{T} = 2 i \underset{k \to \infty}{l i m} {(k M (k | σ_{d}))}_{12} . \end{array}

(3.41)

Proof.The uniqueness of the solution can be guaranteed by Liouville's theorem. For the reflectionless case $V (k) = I$ , it follows from equation (3.36) and the Plemelj formula that one has

\begin{array}{rcl} \begin{array}{rcl} M (k | σ_{d}) & = & I + \sum_{j = 1}^{2 N} \frac{{R e s}_{k = k_{j}} M (k | σ_{d})}{k - k_{j}} + \sum_{j = 1}^{2 N} \frac{{R e s}_{k = k_{j}^{*}} M (k | σ_{d})}{k - k_{j}^{*}} \\ + \sum_{j = 1}^{2 N} \frac{\underset{k = k_{j}}{P_{- 2}} M (k | σ_{d})}{{(k - k_{j})}^{2}} + \frac{\underset{k = k_{j}^{*}}{P_{- 2}} M (k | σ_{d})}{{(k - k_{j}^{*})}^{2}} . \end{array} \end{array}

(3.42)

One can find that

M (k | σ_{d})

has the following formulation of the sum of sparse matrices:

\begin{array}{rcl} \begin{array}{rcl} M (k) & = & I + \sum_{l = 1}^{2 N} (\frac{(β_{l} 0_{3 \times 1})}{k - k_{l}} + \frac{(0_{3 \times 2} \tilde{β_{l}})}{k - k_{l}^{*}} \\ + \frac{(α_{l} 0_{3 \times 1})}{{(k - k_{l})}^{2}} + \frac{(0_{3 \times 2} \tilde{α_{l}})}{{(k - k_{l}^{*})}^{2}}), \end{array} \end{array}

(3.43)

which, together with the residue condition, can further lead to the following equations:

\begin{array}{rcl} (α_{j} 0) = (\begin{array}{cc} 0 & 0 \\ η_{j} & 0 \end{array}) + \sum_{l = 1}^{2 N} (\frac{(\tilde{β_{l}} η_{j} 0)}{k_{j} - k_{l}^{*}} + \frac{(α_{l}^{*} η_{j} 0)}{{(k_{j} - k_{l}^{*})}^{2}}), \end{array}

(3.44)

\begin{array}{rcl} \begin{array}{rcl} (β_{j} 0) & = & (\begin{array}{cc} 0 & 0 \\ η_{j} & 0 \end{array}) + \sum_{l = 1}^{2 N} (\frac{(β_{l}^{*} ζ_{j} 0)}{k_{j} - k_{l}^{*}} + \frac{(α_{l}^{*} ζ_{j} 0)}{{(k_{j} - k_{l}^{*})}^{2}} \\ - \frac{(β_{l}^{*} η_{j} 0)}{{(k_{j} - k_{l}^{*})}^{2}} - 2 \frac{(α_{l}^{*} η_{j} 0)}{{(k_{j} - k_{l}^{*})}^{3}}), \end{array} \end{array}

(3.45)

\begin{array}{rcl} \begin{array}{rcl} (0 α_{j}^{*}) & = & (\begin{array}{cc} 0 & - η_{j}^{†} \\ 0 & 0 \end{array}) + \sum_{l = 1}^{2 N} (\frac{(0 - β_{l} η_{j}^{†})}{k_{j}^{*} - k_{l}} \\ + \frac{(0 - α_{l} η_{j}^{†})}{{(k_{j}^{*} - k_{l})}^{2}}), \end{array} \end{array}

(3.46)

\begin{array}{rcl} \begin{array}{l} (0 β_{j}^{*}) = (\begin{array}{cc} 0 & - η_{j}^{†} \\ 0 & 0 \end{array}) + \sum_{l = 1}^{2 N} (\frac{(0 - β_{l} ζ_{j}^{†})}{k_{j}^{*} - k_{l}} + \frac{(0 - α ζ_{j}^{†})}{k_{j}^{*} - k_{l}} \\ + \frac{(0 β_{l} η_{j}^{†})}{{(k_{j} - k_{l}^{*})}^{2}} + 2 \frac{(0 α_{l} η_{j}^{†})}{{(k_{j} - k_{l}^{*})}^{3}}), \end{array} \end{array}

(3.47)

where

ζ_{j} = B_{j} e^{2 i t θ (k_{j})}, η_{j} = A_{j} e^{2 i t θ (k_{j})}

. Then

α_{l}, β_{l}, \tilde{α_{l}}, \tilde{β_{l}}

can be solved from the above equations.

In what follows, we separate

M_{R H P}^{(2)} (k)

into two parts:

\begin{array}{rcl} \begin{array}{l} M_{R H P}^{(2)} (k) \\ = {\begin{array}{cc} E (k) M^{(o u t)} (k), & k \in C ∖ U_{\pm k_{0}}, \\ E (k) M^{(L C)} (k), M^{(L C)} (k) = M^{(o u t)} (k) M^{(i n)} (k), & k \in U_{\pm k_{0}}, \end{array} \end{array} \end{array}

(3.48)

where M^(out) is used to find the pure solitions outside

U_{\pm k_{0}}

, which is defined in

C

and only admits discrete spectra without a jump. M⁽ⁱⁿ⁾ is defined in

U_{\pm k_{0}}

without discrete spectra, and the model RHP considered by Liu [31]. Moreover, E(k) denotes the error between

M_{R H P}^{(2)} (k)

and M^(out)(k) outside

U_{\pm k_{0}}

Let

\begin{array}{rcl} M^{(o u t)} (k; x, t | σ_{d}^{(o u t)}) = M (k; x, t | σ_{d}) (\begin{array}{cc} δ^{- 1} (k) & 0 \\ 0 & det δ (k) \end{array}), \end{array}

(3.49)

with the scattering data

\begin{array}{rcl} \begin{array}{rcl} σ_{d}^{(o u t)} & = & {(k_{j}, {\tilde{A}}_{j}, {\tilde{B}}_{j}, k_{j} \in K}_{j = 1}^{2 N}, \\ {{\tilde{A}}_{j}, {\tilde{B}}_{j}} & = & {A_{j}, B_{j}} δ^{- 1} (k_{j}) {(det δ (k_{j}))}^{- 1} . \end{array} \end{array}

(3.50)

Then

M^{(o u t)} (k | σ_{d}^{(o u t)})

satisfies the following RH problem:

RHP-5. Find a matrix-valued function

M^{(o u t)} (k; x, t | σ_{d}^{o u t})

without the jump condition solving the following problem:

•	•Analyticity: $M^{(o u t)} (k \| σ_{d}^{o u t})$ is analytic in $C ∖ (Σ^{(2)} \cup K \cup \bar{K})$ .
•	•Asymptotics: $M^{(o u t)} (k \| σ_{d}^{o u t}) \to I$ , k → ∞ ,

M^{(o u t)} (k | σ_{d}^{o u t})

has double poles at k_j and

k_{j}^{*}

with

\begin{array}{rcl} \underset{k = k_{j}}{P_{- 2}} M^{(1)} (k; x, t) = \underset{k \to k_{j}}{l i m} M^{(1)} (k) (\begin{array}{cc} 0 & 0 \\ A_{j} T_{1}^{- 1} T_{2}^{- 1} e^{2 i t θ (k)} & 0 \end{array}), \end{array}

(3.51)

\begin{array}{rcl} \begin{array}{l} {R e s}_{k = k_{j}} M^{(1)} (k; x, t) = \underset{k \to k_{j}}{l i m} M^{(1)} (k) \\ \times (\begin{array}{cc} 0 & 0 \\ (B_{j} T_{1}^{- 1} T_{2}^{- 1} + A_{j} T_{1}^{- 1} (T_{2}^{- 1})^{'}) e^{2 i t θ (k)} & 0 \end{array}) \\ + (M^{(1)})^{'} (k) (\begin{array}{cc} 0 & 0 \\ A_{j} T_{1}^{- 1} T_{2}^{- 1} e^{2 i t θ (k)} & 0 \end{array}), \end{array} \end{array}

(3.52)

\begin{array}{rcl} \underset{k = k_{j}^{*}}{P_{- 2}} M^{(1)} (k; x, t) = \underset{k \to k_{j}^{*}}{l i m} M^{(1)} (\begin{array}{cc} 0 & {\hat{A}}_{j} T_{1} T_{2} e^{2 i t θ} \\ 0 & 0 \end{array}), \end{array}

(3.53)

\begin{array}{rcl} \begin{array}{l} {R e s}_{k = k_{j}^{*}} M^{(1)} (k; x, t) = \underset{k \to k_{j}^{*}}{l i m} M^{(1)} (k) \\ \times (\begin{array}{cc} 0 & ({\hat{B}}_{j} T_{1} T_{2} + {\hat{A}}_{j} T_{2} (T_{1})^{'}) e^{- 2 i t θ (k)} \\ 0 & 0 \end{array}) \\ + (M^{(1)})^{'} (k) (\begin{array}{cc} 0 & {\hat{A}}_{j} T_{1} T_{2} e^{- 2 i t θ (k)} \\ 0 & 0 \end{array}) . \end{array} \end{array}

(3.54)

Proposition 5.RHP-5 has the uniqueness solution, and its potential is equivalent to one of the reflectionless cases of RHP-4, that is

\begin{array}{rcl} \begin{array}{l} q_{s o l} (x, t | σ_{d}^{(o u t)}) = q_{s o l} (x, t | σ_{d}) = {(q (x, t), q^{*} (x, t))}^{T} \\ = 2 i \underset{k \to \infty}{l i m} {(k M (k | σ_{d}))}_{12} . \end{array} \end{array}

(3.55)

Proof.According to the reconstruction formula (2.24), the proof is similar to [32].

We now consider the long-time asymptotic behavior of soliton solutions. Firstly, we define a space-time region

\begin{array}{rcl} D (v_{1}, v_{2}) = {(x, t) \in R^{2} | x = v t, v \in [v_{1}, v_{2}]}, \end{array}

(3.56)

where v₂ ≤ v₁ < 0. Let

\begin{array}{rcl} \begin{array}{rcl} I & = & [- \frac{v_{1}}{4}, - \frac{v_{2}}{4}], \\ K (I) & = & {k_{j} \in K | - \frac{v_{1}}{4} ⩽ K_{j} ⩽ - \frac{v_{2}}{4}}, \\ N (I) & = & | K (I) |, \\ K^{-} (I) & = & {k_{j} \in K | K_{j} < - \frac{v_{1}}{4}}, \\ K^{+} (I) & = & {k_{j} \in K | K_{j} > - \frac{v_{2}}{4}}, \\ K_{j} & = & 3 {R e}^{2} k_{j} - {I m}^{2} k_{j} \\ c_{j} (I) & = & c_{j} δ^{- 1} (k_{j}) e^{\frac{i}{2 π} \int_{- k_{0}}^{k_{0}} \frac{l o g (1 + | γ (ζ) |^{2})}{ζ - k_{j}} d ζ} . \end{array} \end{array}

(3.57)

See figure 3. Then we have the following proposition:

Proposition 6.Given scattering data $σ_{d} = {(k_{j}, A_{j}, B_{j})}_{j = 1}^{2 N}$ and $σ_{d} (I) = {(k_{j}, c_{j} (I)) | k_{j} \in K (I)}$ . At $t \to + \infty$ with $(x, t) \in D (v_{1}, v_{2})$ , we have

\begin{array}{rcl} M (k; x, t | σ_{d}) = [I + O (e^{- 8 μ t})] M^{Δ_{I}} (k; x, t | σ_{d} (I)), \end{array}

(3.58)

where

μ (I) = {min}_{k_{j} \in K / K (I)} {I m k_{j} \cdot d i s t (3 {R e}^{2} k_{j} - {I m}^{2} k_{j}, I)}

Proof.The proof is similar to [13, 32].

Corollary 1.Suppose that q_sol is the soliton solution of the SS equation corresponding to its scattering data $σ_{d} = {(k_{j}, A_{j}, B_{j})}_{j = 1}^{2 N}$ , then one has

\begin{array}{rcl} q_{s o l} (x, t | σ_{d}^{(o u t)}) = q_{s o l} (x, t | σ_{d} (I)) + O (e^{- 8 μ t}), t \to + \infty, \end{array}

(3.59)

where

q_{s o l} (x, t | σ_{d} (I))

is the soliton solution corresponding to the scattering data

σ_{d} (I)

of the SS equation.

3.2.2. The solvable local RH problem

RHP-6. Find a matrix-valued function M⁽ⁱⁿ⁾(k; x, t) which satisfies

•

•Analyticity: M⁽ⁱⁿ⁾(k; x, t) is analytical in $C ∖ Σ^{(2)}$ with symmetry: M⁽ⁱⁿ⁾(k) = ϱM^(in)*(− k^*)ϱ.

•

•Jump condition: M⁽ⁱⁿ⁾(k; x, t) has the jump condition

\begin{array}{rcl} M_{+}^{(i n)} (k) = M_{-}^{(i n)} (k) V^{(i n)} (k), k \in Σ^{(2)} . \end{array}

(3.60)

where the jump matrix V⁽ⁱⁿ⁾(k) = V⁽²⁾(k) is given by equation (3.27). See figure 4.

•

•Asymptotics: M⁽ⁱⁿ⁾(k) → I, k → ∞ .

RHP-6 is a solvable model for the SS equation. Here, we mainly adopt the final results for solving the model RHP (see [31] for more details), whose solution has the asymptotics:

\begin{array}{rcl} \begin{array}{rcl} M^{(i n)} (k) & = & I - \frac{1}{\sqrt{48 {t k}_{0}} (k + k_{0})} M_{1}^{(i n)} \\ + \frac{1}{\sqrt{48 {t k}_{0}} (k - k_{0})} ϱ {(M_{1}^{(i n)})}^{*} ϱ + O (\frac{l o g t}{t}), \end{array} \end{array}

(3.61)

with ∥M⁽ⁱⁿ⁾∥_∞ ≲ 1, where

\begin{array}{rcl} M_{1}^{(i n)} = (\begin{array}{cc} 0 & i ϖ^{- 2} β_{12} \\ - i ϖ^{2} β_{21} & 0 \end{array}), \end{array}

(3.62)

with

\begin{array}{rcl} \begin{array}{rcl} ϖ & = & {(192 τ)}^{\frac{i ν}{2}} e^{X (- k_{0}) - 8 i τ}, \\ β_{12} & = & - β_{21}^{†} = \frac{ν Γ (- i ν) e^{\frac{π (2 ν - i)}{4}}}{\sqrt{2 π}} σ_{2} γ^{T} (k_{0}) . \end{array} \end{array}

(3.63)

According to RHP-5 and RHP-6, one has the solvable local model RHP with M^(LC)(k) = M^(out)(k)M⁽ⁱⁿ⁾(k) inside

U_{\pm k_{0}}

which is a bounded function in

U_{\pm k_{0}}

and has the same jump condition as

M_{R H P}^{(2)} (k)

3.2.3. A small norm RH problem

In this section, we mainly consider the small norm RHP corresponding to the error matrix function E(k) given by equation (3.48). Firstly, according to the definition of

M_{R H P}^{(2)} (k)

and M^(LC)(k), we can obtain that E(k) satisfies the following RHP:

RHP-7. Find a matrix-valued function E(k) solving

•

•Analyticity: E(k) is continuous in $C ∖ Σ^{(E)}$ , where $Σ^{(E)} = \partial U_{\pm k_{0}} \cup (Σ^{(E)} ∖ U_{\pm k_{0}}) .$

•

•Jump condition: E(k) has the following jump condition (see figure 5)

\begin{array}{rcl} E_{+} (k) = E_{-} (k) V^{(E)} (k), k \in Σ^{(E)}, \end{array}

(3.64)

where matrix V^(E)(k) is defined by

\begin{array}{rcl} V^{(E)} (k) = {\begin{array}{cc} M^{(o u t)} (k) V^{(2)} (k) M^{(o u t)} {(k)}^{- 1}, & k \in Σ^{(2)} ∖ U_{\pm k_{0}}, \\ M^{(o u t)} (k) M^{(i n)} (k) M^{(o u t)} {(k)}^{- 1}, & k \in \partial U_{\pm k_{0}} . \end{array} \end{array}

(3.65)

•

•Asymptotic behaviors: E(k) → I, k → ∞ .

Proposition 7.The jump matrix $V^{(E)} (k)$ has the following uniform estimate

\begin{array}{rcl} | V^{(E)} (k) - I | = {\begin{cases} O (e^{- 6 k_{0} ρ^{2} t}), & k \in Σ_{j}^{\pm} ∖ U_{\pm k_{0}}, j = 1, 2, \\ O (e^{- 8 k_{0}^{2} ρ t}), & k \in Σ_{j}^{\pm} ∖ U_{\pm k_{0}}, j = 3, 4, \\ O (e^{- 14 k_{0}^{3} \tan^{3} (π / 12) t}), & k \in [\pm i k_{0}, \pm i k_{0} \tan (\frac{π}{12})], \\ O (t^{- 1 / 2}), & k \in \partial U_{\pm k_{0}}, \\ 0, & k \in [- i k_{0} \tan (\frac{π}{12}), i k_{0} \tan (\frac{π}{12}) . \end{cases} \end{array}

(3.66)

Proof.The proof can be seen in [32].

Based on the Beals–Coifman theorem, we can construct the solution of RHP-7 in the form

\begin{array}{rcl} E (k) = I + \frac{1}{2 π i} \int_{Σ^{(E)}} \frac{κ_{E} (ξ) [V^{(E)} (ξ) - I]}{ξ - k} d ξ, \end{array}

(3.67)

where κ_E ∈ L²(Σ^(E)) satisfies

(I - C_{ω_{E}}) κ_{E} = I

\begin{array}{rcl} \begin{array}{rcl} ω_{E} & = & {(ω_{E})}_{+} + {(ω_{E})}_{-} = V^{(E)} - I, {(ω_{E})}_{-} = 0, \\ {(ω_{E})}_{+} = V^{(E)} - I, \\ C_{ω_{E}} g & = & C_{-} (g {(ω_{E})}_{+}) + C_{+} (g {(ω_{E})}_{-}) = C_{-} (g (V^{(E)} - I)) \end{array} \end{array}

(3.68)

with C₋ denoting the Cauchy projection operator

\begin{array}{rcl} C_{-} g (k) = \underset{ζ \to k \in Σ^{(E)}}{l i m} \frac{1}{2 π i} \int_{Σ^{(E)}} \frac{g (ξ)}{ξ - ζ} d ξ, \end{array}

(3.69)

and

∥ C_{-} ∥_{L^{2}}

is a finite value.

Since

\begin{array}{rcl} ∥ C_{ω_{E}} ∥_{L^{2} (Σ^{(E)})} ≲∥ C ∥_{L^{2} (Σ^{(E)})} ∥ V^{(E)} - I ∥_{L^{\infty} (Σ^{E})} ≲ O (t^{- 1 / 2}), \end{array}

(3.70)

the matrix function κ_E exists and is unique, and the solution E(k) of RHP-7 exists and is unique.

Proposition 8. $V^{(E)}$ and $κ_{E}$ admit the following important estimates

\begin{array}{rcl} \begin{array}{l} ∥ κ_{E} - I ∥_{L^{2} (Σ^{(E)})} = O (t^{- 1 / 2}), \\ ∥ V^{(E)} - I ∥_{L^{p}} = O (t^{- 1 / 2}), p \in [1, + \infty), k ⩾ 0. \end{array} \end{array}

(3.71)

Proposition 9.The matrix function E(k) has the following asymptotics

\begin{array}{rcl} E (k; x, t) = I + \frac{E_{1} (x, t)}{k} + O (k^{- 2}), k \to \infty, \end{array}

(3.72)

where

\begin{array}{rcl} E_{1} (x, t) = \frac{i}{2 π} \int_{Σ^{(E)}} κ_{E} (ξ) (V^{E} - I) d ξ . \end{array}

(3.73)

Moreover,

E_{1} (x, t)

is given by

\begin{array}{rcl} \begin{array}{l} E_{1} (x, t) = \frac{1}{\sqrt{48 {t k}_{0}}} M^{(o u t)} (k_{0}) M_{1}^{(i n)} (k_{0}) M^{(o u t) - 1} (k_{0}) \\ + \frac{1}{\sqrt{48 {t k}_{0}}} M^{(o u t)} (- k_{0}) M_{1}^{B_{0}} (- k_{0}) M^{(o u t) - 1} (- k_{0}) \\ + O (t^{- 1} l o g t) . \end{array} \end{array}

(3.74)

Proof.The proof is similar as [32].

3.3. Analysis on a pure $\bar{\partial}$ -problem

Here we consider the pure

\bar{\partial}

-problem which is obtained by removing the pure RHP part with

\bar{\partial} R^{(2)} = 0

. Let

\begin{array}{rcl} M^{(3)} (k) = M^{(2)} (k) M_{R H P}^{(2)} {(k)}^{- 1} . \end{array}

(3.75)

Then we know that M⁽³⁾ is continuous and has no jumps in the complex plane, and solves a pure

\bar{\partial}

-problem. Pure

\bar{\partial}

-problem. Find a matrix-valued function M⁽³⁾(k, x, t) solving

•	•Continuity: M⁽³⁾(k) is continuous in $C ∖ Σ^{(2)} .$
•	•Jump condition: $\bar{\partial} M^{(3)} (k) = M^{(3)} (k) W^{(3)} (k)$ , $k \in C$ , where $W^{(3)} = M_{R H P}^{(2)} (k) \bar{\partial} R^{(2)} (k) M_{R H P}^{(2)} {(k)}^{- 1} .$
•	•Asymptotic behaviors: M⁽³⁾(k) → I, k → ∞ .

The solution of the above pure

\bar{\partial}

-problem can be given by the following integral equation

\begin{array}{rcl} (I - F) M^{(3)} (k) = I, \end{array}

(3.76)

where F is the Cauchy operator defined as

\begin{array}{rcl} F [f] (k) = - \frac{1}{π} \iint_{C} \frac{f (ξ) W^{(3)} (ξ)}{ξ - k} d A (ξ) \end{array}

(3.77)

with dA(ξ) being the Lebesgue measure.

Proposition 10.For large time t, there exists the estimate for F:

\begin{array}{rcl} ∥ F ∥_{L^{\infty} \to L^{\infty}} ≲ {(k_{0} t)}^{- 1 / 4}, \end{array}

(3.78)

which implies that the operator

{(I - F)}^{- 1}

is invertible and the solution of the pure

\bar{\partial}

-problem exists and is unique.

M⁽³⁾(k) has the following asymptotic expansion

\begin{array}{rcl} \begin{array}{rcl} M^{(3)} & = & I + \frac{M_{1}^{(3)}}{k} + O (\frac{1}{k^{2}}), \\ M_{1}^{(3)} & = & \frac{1}{π} \iint_{C} M^{(3)} (ξ) W^{(3)} (ξ) d A (ξ), \\ k & \to & \infty . \end{array} \end{array}

(3.79)

According to proposition 10 and the asympotics of

M_{1}^{(3)}

, one has

Proposition 11.For large time t, there exists the estimate for $M_{1}^{(3)}$

\begin{array}{rcl} | M_{1}^{(3)} | ≲ {(k_{0} t)}^{- 3 / 4} . \end{array}

(3.80)

4. Long-time asymptotics in the region $x < 0, | x / t | = O (1)$ and $x > 0, | x / t | = O (1)$

Based on the above discussions, our main result is summarized as follows:

Theorem 1.Let $σ_{d} = {(k_{j}, A_{j}, B_{j}), k_{j} \in K}_{j = 1}^{2 N}$ denote the scattering data generated by initial data $q_{0} (x) \in S (R)$ with the second-order discrete spectra. For fixed $v_{2} ⩽ v_{1} \in R^{-}$ , define $I = [- v_{1} / 4, - v_{2} / 4]$ and a space-time cone $D (v_{1}, v_{2})$ for variables x and t. Let $q_{s o l} (x, t, σ_{d} (I))$ be the $N (I)$ solution corresponding to the modified scattering data $σ_{d} (I) = {(k_{j}, c_{j} (I)), k_{j} \in K (I)}$ . When $x < 0$ , as $t \to + \infty$ with $(x, t) \in D (v_{1}, v_{2})$ , we have the long-time asymptotics of the SS equation

\begin{array}{rcl} q (x, t) = q_{s o l} (x, t | σ_{d} (I)) + {p t}^{- 1 / 2} + O (t^{- 3 / 4}), \end{array}

(4.1)

where

\begin{array}{rcl} \begin{array}{rcl} p & = & \frac{1}{\sqrt{48 k_{0}}} (M^{(o u t)} (k_{0}) M_{1}^{(i n)} (k_{0}) M^{(o u t) - 1} (k_{0}) \\ {- M^{(o u t)} (- k_{0}) ϱ {(M_{1}^{(i n)} (- k_{0}))}^{*} ϱ M^{(o u t) - 1} (- k_{0}))}_{12} . \end{array} \end{array}

(4.2)

Similarly, when

x > 0

, as

t \to + \infty

with

(x, t) \in D (v_{1}, v_{2})

, we have

\begin{array}{rcl} q (x, t) = q_{s o l} (x, t | σ_{d} (I)) + O (t^{- 1}) . \end{array}

(4.3)

Proof.Based on a series of transformations (3.12), (3.25), (3.34) and (3.48), we find

\begin{array}{rcl} M (k) = M^{(3)} (k) E (k) M^{(o u t)} (k) R^{(2) - 1} (k) T^{- 1} (k) . \end{array}

In particular, by considering

k \to \infty

along the imaginary axis (i.e. in

Ω_{5}, Ω_{6}

), we have

\begin{array}{rcl} \begin{array}{l} M = (I + \frac{M_{1}^{(3)}}{k} + . . .) (I + \frac{E_{1}}{k} + . . .) (I + \frac{M_{1}^{(o u t)}}{k} + . . .) \\ \times (I + \frac{T_{1}}{k} + . . .) = I + \frac{M_{1}}{k} + . . ., \end{array} \end{array}

which generates

\begin{array}{rcl} M_{1} = M_{1}^{(o u t)} + E_{1} + M_{1}^{(3)} + T_{1} . \end{array}

(4.4)

According to the reconstruction formula (2.24) and proposition 11, the following estimate holds

\begin{array}{rcl} q (x, t) = 2 i {(M_{1}^{(o u t)})}_{12} + 2 i {(E_{1})}_{12} + O (t^{- 3 / 4}) . \end{array}

(4.5)

Notice that

\begin{array}{rcl} 2 i {(M_{1}^{(o u t)})}_{12} = q_{s o l} (x, t | σ_{d}^{(o u t)}), \end{array}

(4.6)

which, together with proposition 9, yields

\begin{array}{rcl} {(E_{1})}_{12} = {p t}^{- 1 / 2} + O (t^{- 1} l o g t), \end{array}

(4.7)

where p is given by equation (4.2). Substituting equations (4.6) and (4.7) into (4.5) yields

\begin{array}{rcl} q (x, t) = q_{s o l} (x, t | σ_{d}^{(o u t)}) + {p t}^{- 1 / 2} + O (t^{- 3 / 4}) . \end{array}

(4.8)

Based on equation (3.59), we find the final asymptotic expression (4.1) with

(x, t) \in D (v_{1}, v_{2})

Remark 1.Though the large-time asymptotics of the potential given by equation (4.1) has the same form as in [32], they have different meanings. In (4.1), $q_{s o l} (x, t | σ_{d} (I))$ denotes the soliton solutions generated by the double poles of the scattering data of the spectral problem, while it denotes the simple poles case in [32].

Remark 2.Theorem 1 did not consider the Painlevé asymptotics in the Painlevé region, in which the main term of the potential has no soliton solution. Thus, one need not consider the order of the discrete spectra such that the formula of the corresponding asymptotic behavior is the same as the one in [32].

References

List( Publishing order | Descend order by publishing year | Descend order by cited within ) Chart analysis

1	Deift P A, Zhou X 1993 A steepest descent method for oscillatory Riemann-Hilbert problems Ann. Math. 137 295 368 https://doi.org/10.2307/2946540 Cited in this article [1]

2	Cheng P J, Venakides S, Zhou X 1999 Long-time asymptotics for the pure radiation solution of the sine-Gordon equation Commun. PDE 24 1195 1262 https://doi.org/10.1080/03605309908821464 Cited in this article [1]

3	Grunert K, Teschl G 2009 Long-time asymptotics for the Korteweg-de Vries equation via nonlinear steepest descent Math. Phys. Anal. Geom. 12 287 324 https://doi.org/10.1007/s11040-009-9062-2 Cited in this article [1]

4	Boutet de Monvel A, Its A, Shepelsky D 2010 Painlevé-type asymptotics for the Camassa-Holm equation SIAM J. Math. Anal. 42 1854 1873 https://doi.org/10.1137/090772976 Cited in this article [1]

5	Boutet de Monvel A, Shepelsky D, Zielinski L 2017 The short pulse equation by a Riemann-Hilbert approach Lett. Math. Phys. 107 1345 1373 https://doi.org/10.1007/s11005-017-0945-z Cited in this article [1]

6	Xu J 2018 Long-time asymptotics for the short pulse equation J. Differ. Equ. 265 3494 3532 https://doi.org/10.1016/j.jde.2018.05.009 Cited in this article [1]

7	Xu J, Fan E 2015 Long-time asymptotics for the Fokas-Lenells equation with decaying initial value problem: without solitons J. Differ. Equ. 259 1098 1148 https://doi.org/10.1016/j.jde.2015.02.046 Cited in this article [1]

8	Liu N, Guo B 2021 Painlevé-type asymptotics of an extended modified KdV equation in transition regions J. Differ. Equ. 280 203 235 https://doi.org/10.1016/j.jde.2021.01.020 Cited in this article [1]

9	Liu N, Guo B, Wang D, Wang Y 2019 Long-time asymptotic behavior for an extended modified Korteweg-de Vries equation Commun. Math. Sci. 17 1877 1913 https://doi.org/10.4310/CMS.2019.v17.n7.a6 Cited in this article [1]

McLaughlin

K T R

, Miller

P D

2006 The

\bar{\partial}

steepest descent method and the asymptotic behavior of polynomials orthogonal on the unit circle with fixed and exponentially varying non-analytic weights Int. Math. Res. Not. 2006 48673

https://doi.org/10.1155/IMRP/2006/48673

Cited in this article [1]

11	McLaughlin K T R, Miller P D 2008 The $\bar{\partial}$ steepest descent method for orthogonal polynomials on the real line with varying weights Int. Math. Res. Not. 2008 075 https://doi.org/10.1093/imrn/rnn075 Cited in this article [1]

12	Borghese M, Jenkins R, McLaughlin K T-R 2018 Long time asymptotic behavior of the focusing nonlinear Schrödinger equation Ann. Inst. Henri Poincaré Anal. Non Linéaire 35 887 920 https://doi.org/10.1016/j.anihpc.2017.08.006 Cited in this article [3]

13	Wang Z, Chen M, Fan E 2022 Long time asymptotics for the focusing nonlinear Schrödinger equation in the solitonic region with the presence of high-order discrete spectrum J. Math. Anal. Appl. 505 125635 https://doi.org/10.1016/j.jmaa.2021.125635 Cited in this article [3]

14	Wang Z, Fan E 2023 The defocusing nonlinear Schrödinger equation with a nonzero background: Painlevé asymptotics in two transition regions Commun. Math. Phys. 402 2879 2930 https://doi.org/10.1007/s00220-023-04787-6 Cited in this article [1]

15	Cuccagna S, Jenkins R 2016 On the asymptotic stability of n-soliton solutions of the defocusing nonlinear schrödinger equation Commun. Math. Phys. 343 921 969 https://doi.org/10.1007/s00220-016-2617-8 Cited in this article [1]

16	Jenkins R, Liu J, Perry P, Sulem C 2018 Soliton resolution for the derivative nonlinear Schrödinger equation Comm. Math. Phys. 363 1003 1049 https://doi.org/10.1007/s00220-018-3138-4 Cited in this article [1]

17	Chen G, Liu J 2001 Soliton resolution for the focusing modified KdV equation Ann. Inst. Henri Poincaré Anal. Non Linéaire 38 2005 2071 Cited in this article [1]

18	Xu T, Zhang Z, Fan E 2023 On the Cauchy problem of defocusing mKdV equation with finite density initial data: Long time asymptotics in soliton-less regions J. Differ. Equ. 372 55 122 https://doi.org/10.1016/j.jde.2023.06.038 Cited in this article [1]

19	Liu N, Chen M, Guo B 2021 Long-time asymptotic behavior of the fifth-order modified KdV equation in low regularity spaces Stud. Appl. Math. 147 230 299 https://doi.org/10.1111/sapm.12379 Cited in this article [1]

20	Li Z, Tian S, Yang J, Fan E 2022 Soliton resolution for the complex short pulse equation with weighted Sobolev initial data in space-time solitonic regions J. Differ. Equ. 329 31 88 https://doi.org/10.1016/j.jde.2022.05.003 Cited in this article [1]

21	Yang Y, Fan E 2022 On the long-time asymptotics of the modified Camassa-Holm equation in space-time solitonic regions Adv. Math. 402 108340 https://doi.org/10.1016/j.aim.2022.108340 Cited in this article [1]

22	Yang Y, Fan E 2023 Soliton resolution and large time behavior of solutions to the Cauchy problem for the Novikov equation with a nonzero background Adv. Math. 426 109088 https://doi.org/10.1016/j.aim.2023.109088 Cited in this article [1]

23	Sasa N, Satsuma J 1991 New-type of soliton solutions for a high-order nonlinear Schrödinger equation J. Phys. Soc. Japan 60 409 417 https://doi.org/10.1143/JPSJ.60.409 Cited in this article [1]

24	Tsuchida T, Wadati M 1998 The coupled modified Korteweg-de Vries equations J. Phys. Soc. Japan 67 1175 1187 https://doi.org/10.1143/JPSJ.67.1175 Cited in this article [1]

25	Yang J 2010 Nonlinear Waves in Integrable and Nonintegrable Systems (SIAM) Cited in this article [1]

26	Gilson C, Hietarinta J, Nimmo J, Ohta Y 2003 Sasa–Satsuma high-order nonlinear Schrödinger equation and its bilinearization and multisoliton solutions Phys. Rev. E 68 016614 https://doi.org/10.1103/PhysRevE.68.016614 Cited in this article [1]

27	Kim J, Park Q H, Shin H J 1998 Conservation laws in high-order nonlinear Schrödinger equations Phys. Rev. E 58 6746 6751 https://doi.org/10.1103/PhysRevE.58.6746

28	Sergyeyev A, Demskoi D 2007 Sasa–Satsuma (complex modified Korteweg-de Vries II) and the complex sine-Gordon II equation revisited: recursion operators, nonlocal symmetries, and more J. Math. Phys. 48 042702 https://doi.org/10.1063/1.2710552

29	Akhmediev N, Soto-Crespo J M, Devine N, Hoffmann N P 2015 Rogue wave spectra of the Sasa-Satsuma equation Phys. D 294 37 42 https://doi.org/10.1016/j.physd.2014.11.006

30	Wu L H, Geng X G, He G L 2016 Algebro-geometric solutions to the Manakov hierarchy Appl. Anal. 95 769 800 https://doi.org/10.1080/00036811.2015.1031220 Cited in this article [1]

31	Liu H, Geng X G, Xue B 2018 The Deift–Zhou steepest descent method to long time asymptotics for the Sasa–Satsuma equation J. Differ. Equ. 265 5984 6008 https://doi.org/10.1016/j.jde.2018.07.026 Cited in this article [3]

32	Xun W, Fan E 2022 Long time and Painlevé-type asymptotics for the Sasa–Satsuma equation in solitonic space time regions J. Differ. Equ. 329 89 130 https://doi.org/10.1016/j.jde.2022.05.006 Cited in this article [10]

RIGHTS & PERMISSIONS

PDF(652 KB)

Collection(s)

Mathematical Physics

Accesses

Citation

Detail

Sections

Recommended

Abstract
Key words
Cite this article
1. Introduction
2. The direct scattering problem
2.1. Jost solutions of the Lax pair and scattering data
2.2. The Riemann–Hilbert problem with higher-order poles
3. The mixed $\bar{\partial}$ -RH problem and its decomposition
3.1. Two factorizations of jump matrix V(k)
Figure 1. The signature table of $R e i θ (k) = 4 I m k ({I m}^{2} k - 3 {R e}^{2} k + 3 k_{0}^{2})$ with ±k0 being phase points.
Figure 2. Deformation of the jump countor from $R$ to Σ(2).
Figure 3. Pole distribution. The red, green and yellow points generate the breather solutions. Moreover, the red points lie in the region $K^{+} (I)$ , the green points lie in the region $K (I)$ , and the yellow points on the line $R e i θ (k) = 0$ .
Figure 4. The jump contour for the jump matrix V(in).
Figure 5. The jump contour Σ(E) for the error function E(k).
3.2. Analysis on a pure RH problem
3.2.1. Soliton solutions corresponding to discrete spectra
3.2.2. The solvable local RH problem
3.2.3. A small norm RH problem
3.3. Analysis on a pure $\bar{\partial}$ -problem
4. Long-time asymptotics in the region $x < 0, | x / t | = O (1)$ and $x > 0, | x / t | = O (1)$
References
RIGHTS & PERMISSIONS

Received	Revised	Accepted	Published
2024-02-05	2024-03-18	2024-03-21	2024-05-13
Issue Date
2024-05-13

Please choose a citation manager

Content to export

Abstract

Key words

Cite this article

1. Introduction

2. The direct scattering problem

2.1. Jost solutions of the Lax pair and scattering data

2.2. The Riemann–Hilbert problem with higher-order poles

3. The mixed $\bar{\partial}$ -RH problem and its decomposition

3.1. Two factorizations of jump matrix V(k)

Figure 1. The signature table of $R e i θ (k) = 4 I m k ({I m}^{2} k - 3 {R e}^{2} k + 3 k_{0}^{2})$ with ±k₀ being phase points.

Figure 2. Deformation of the jump countor from $R$ to Σ⁽²⁾.

Figure 3. Pole distribution. The red, green and yellow points generate the breather solutions. Moreover, the red points lie in the region $K^{+} (I)$ , the green points lie in the region $K (I)$ , and the yellow points on the line $R e i θ (k) = 0$ .

Figure 4. The jump contour for the jump matrix V⁽ⁱⁿ⁾.

Figure 5. The jump contour Σ^(E) for the error function E(k).

3.2. Analysis on a pure RH problem

3.2.1. Soliton solutions corresponding to discrete spectra

3.2.2. The solvable local RH problem

3.2.3. A small norm RH problem

3.3. Analysis on a pure $\bar{\partial}$ -problem

4. Long-time asymptotics in the region $x < 0, | x / t | = O (1)$ and $x > 0, | x / t | = O (1)$

{{custom_sec.title}}

{{custom_sec.title}}

References

{{custom_fnGroup.title_en}}

Footnotes

RIGHTS & PERMISSIONS

模态框（Modal）标题

Please choose a citation manager

Content to export

Abstract

Key words

Cite this article

1. Introduction

2. The direct scattering problem

2.1. Jost solutions of the Lax pair and scattering data

2.2. The Riemann–Hilbert problem with higher-order poles

3. The mixed ∂¯-RH problem and its decomposition

3.1. Two factorizations of jump matrix V(k)

Figure 1. The signature table of Reiθ(k)=4Imk(Im2k−3Re2k+3k02) with ±k0 being phase points.

Figure 2. Deformation of the jump countor from R to Σ(2).

Figure 3. Pole distribution. The red, green and yellow points generate the breather solutions. Moreover, the red points lie in the region K+(I), the green points lie in the region K(I), and the yellow points on the line Reiθ(k)=0.

Figure 4. The jump contour for the jump matrix V(in).

Figure 5. The jump contour Σ(E) for the error function E(k).

3.2. Analysis on a pure RH problem

3.2.1. Soliton solutions corresponding to discrete spectra

3.2.2. The solvable local RH problem

3.2.3. A small norm RH problem

3.3. Analysis on a pure ∂¯-problem

4. Long-time asymptotics in the region x<0,|x/t|=O(1) and x>0,|x/t|=O(1)

{{custom_sec.title}}

{{custom_sec.title}}

References

{{custom_fnGroup.title_en}}

Footnotes

RIGHTS & PERMISSIONS

3. The mixed $\bar{\partial}$ -RH problem and its decomposition

Figure 1. The signature table of $R e i θ (k) = 4 I m k ({I m}^{2} k - 3 {R e}^{2} k + 3 k_{0}^{2})$ with ±k₀ being phase points.

Figure 2. Deformation of the jump countor from $R$ to Σ⁽²⁾.

Figure 3. Pole distribution. The red, green and yellow points generate the breather solutions. Moreover, the red points lie in the region $K^{+} (I)$ , the green points lie in the region $K (I)$ , and the yellow points on the line $R e i θ (k) = 0$ .

Figure 4. The jump contour for the jump matrix V⁽ⁱⁿ⁾.

Figure 5. The jump contour Σ^(E) for the error function E(k).

3.3. Analysis on a pure $\bar{\partial}$ -problem

4. Long-time asymptotics in the region $x < 0, | x / t | = O (1)$ and $x > 0, | x / t | = O (1)$