Step-Initial Function to the MKdV Equation: Hyper-Elliptic Long-Time Asymptotics of the Solution

Step-Initial Function to the MKdV Equation: Hyper-Elliptic Long-Time Asymptotics of the Solution

The modified Korteveg{de Vries equation on the line is considered. The initial function is a discontinuous and piece-wise constant step function, i.e. q(x, 0) = cr for x ≥ 0 and q(x, 0) = cl for x < 0, where cl, cr are real numbers which satisfy cl > cr > 0. The goal of this paper is to stu...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Datum:2012
Hauptverfasser: Kotlyarov, V., Minakov, A.
Format: Artikel
Sprache:English
Veröffentlicht: Фізико-технічний інститут низьких температур ім. Б.І. Вєркіна НАН України 2012
Schriftenreihe:Журнал математической физики, анализа, геометрии
Online Zugang:http://dspace.nbuv.gov.ua/handle/123456789/106707
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Назва журналу:Digital Library of Periodicals of National Academy of Sciences of Ukraine
Zitieren:Step-Initial Function to the MKdV Equation: Hyper-Elliptic Long-Time Asymptotics of the Solution / V. Kotlyarov, A. Minakov // Журнал математической физики, анализа, геометрии. — 2012. — Т. 8, № 1. — С. 38-62. — Бібліогр.: 49 назв. — англ.

Institution

Digital Library of Periodicals of National Academy of Sciences of Ukraine
id irk-123456789-106707
record_format dspace
spelling irk-123456789-1067072016-10-04T03:02:16Z Step-Initial Function to the MKdV Equation: Hyper-Elliptic Long-Time Asymptotics of the Solution Kotlyarov, V. Minakov, A. The modified Korteveg{de Vries equation on the line is considered. The initial function is a discontinuous and piece-wise constant step function, i.e. q(x, 0) = cr for x ≥ 0 and q(x, 0) = cl for x < 0, where cl, cr are real numbers which satisfy cl > cr > 0. The goal of this paper is to study the asymptotic behavior of the solution of the initial-value problem as t → ∞. Рассматривается модифицированное уравнение КдФ на всей прямой с начальным условием типа ступеньки, которая равна константе cl при x < 0 и другой константе cr при x > 0. При этом выполняется условие cl > cr > 0, что обеспечивает режим "гидродинамической волны сжатия" при t → ∞. Цель статьи - изучение асимптотического поведения решения начально-краевой задачи, когда t → ∞. 2012 Article Step-Initial Function to the MKdV Equation: Hyper-Elliptic Long-Time Asymptotics of the Solution / V. Kotlyarov, A. Minakov // Журнал математической физики, анализа, геометрии. — 2012. — Т. 8, № 1. — С. 38-62. — Бібліогр.: 49 назв. — англ. 1812-9471 http://dspace.nbuv.gov.ua/handle/123456789/106707 en Журнал математической физики, анализа, геометрии Фізико-технічний інститут низьких температур ім. Б.І. Вєркіна НАН України
institution Digital Library of Periodicals of National Academy of Sciences of Ukraine
collection DSpace DC
language English
description The modified Korteveg{de Vries equation on the line is considered. The initial function is a discontinuous and piece-wise constant step function, i.e. q(x, 0) = cr for x ≥ 0 and q(x, 0) = cl for x < 0, where cl, cr are real numbers which satisfy cl > cr > 0. The goal of this paper is to study the asymptotic behavior of the solution of the initial-value problem as t → ∞.
format Article
author Kotlyarov, V.
Minakov, A.
spellingShingle Kotlyarov, V.
Minakov, A.
Step-Initial Function to the MKdV Equation: Hyper-Elliptic Long-Time Asymptotics of the Solution
Журнал математической физики, анализа, геометрии
author_facet Kotlyarov, V.
Minakov, A.
author_sort Kotlyarov, V.
title Step-Initial Function to the MKdV Equation: Hyper-Elliptic Long-Time Asymptotics of the Solution
title_short Step-Initial Function to the MKdV Equation: Hyper-Elliptic Long-Time Asymptotics of the Solution
title_full Step-Initial Function to the MKdV Equation: Hyper-Elliptic Long-Time Asymptotics of the Solution
title_fullStr Step-Initial Function to the MKdV Equation: Hyper-Elliptic Long-Time Asymptotics of the Solution
title_full_unstemmed Step-Initial Function to the MKdV Equation: Hyper-Elliptic Long-Time Asymptotics of the Solution
title_sort step-initial function to the mkdv equation: hyper-elliptic long-time asymptotics of the solution
publisher Фізико-технічний інститут низьких температур ім. Б.І. Вєркіна НАН України
publishDate 2012
url http://dspace.nbuv.gov.ua/handle/123456789/106707
citation_txt Step-Initial Function to the MKdV Equation: Hyper-Elliptic Long-Time Asymptotics of the Solution / V. Kotlyarov, A. Minakov // Журнал математической физики, анализа, геометрии. — 2012. — Т. 8, № 1. — С. 38-62. — Бібліогр.: 49 назв. — англ.
series Журнал математической физики, анализа, геометрии
work_keys_str_mv AT kotlyarovv stepinitialfunctiontothemkdvequationhyperellipticlongtimeasymptoticsofthesolution
AT minakova stepinitialfunctiontothemkdvequationhyperellipticlongtimeasymptoticsofthesolution
first_indexed 2025-07-07T18:52:53Z
last_indexed 2025-07-07T18:52:53Z
_version_ 1837015366920830976
fulltext Journal of Mathematical Physics, Analysis, Geometry 2012, vol. 8, No. 1, pp. 38–62 Step-Initial Function to the MKdV Equation: Hyper-Elliptic Long-Time Asymptotics of the Solution V. Kotlyarov and A. Minakov Mathematical division, B.I. Verkin Institute for Low Temperature Physics and Engineering 47 Lenin Avenue, 61103 Kharkiv, Ukraine E-mail: kotlyarov@ilt.kharkov.ua minakov@ilt.kharkov.ua Received November 7, 2011 The modified Korteveg–de Vries equation on the line is considered. The initial function is a discontinuous and piece-wise constant step function, i.e. q(x, 0) = cr for x ≥ 0 and q(x, 0) = cl for x < 0, where cl, cr are real numbers which satisfy cl > cr > 0. The goal of this paper is to study the asymptotic behavior of the solution of the initial-value problem as t → ∞. Using the steepest descent method we deform the original oscillatory matrix Riemann–Hilbert problem to explicitly solvable model forms and show that the solution of the initial-value problem has different asymptotic behavior in different regions of the xt plane. In the regions x < −6c2 l t + 12c2 rt and x > 4c2 l t + 2c2 rt the main term of asymptotics of the solution is equal to cl and cr, respectively. In the region (−6c2 l + 12c2 r)t < x < (4c2 l + 2c2 r)t the asymptotics of the solution takes the form of a modulated hyper-elliptic wave generated by an algebraic curve of genus 2. Key words: modified Korteweg–de Vries equation, step-like initial value problem, Riemann–Hilbert problem, steepest descent method, modulated hyper-elliptic wave. Mathematics Subject Classification 2000: 35Q15, 35B40. 1. Introduction The inverse scattering transform method (IST) [1–3] used for initial-value problems for nonlinear integrable equations is proved to be very successful. It allows to obtain a large number of very interesting results in various areas of mathematics and physics. The IST method was further developed by P. Deift and X. Zhou [4–6]. They proposed to use the steepest descent method for solving the oscillatory matrix Riemann–Hilbert problems. This method appeared to give c© V. Kotlyarov and A. Minakov, 2012 Step-Initial Function to the mKdV Equation a nice possibility to study asymptotic behavior of the solutions of initial-value problems as well as many other problems of the theory of completely integrable nonlinear equations (sf. [7–26]), random matrix models, orthogonal polynomials and integrable statistical mechanics [27–29] without a priori assumptions. The initial value problems for nonlinear integrable equations with step-like initial functions have a very long history. More about these problems can be found in [2, 30–46], and also in the references therein. Most results were obtained for the initial-value problems associated with self-adjoint Lax operators. First the step-like problems with non self-adjoint Lax operators were considered by Bikbaev [40, 41] and later by Novokshenov [43]. In their papers, the main attention was paid to the studying of the complex Whitham deformations, which allowed them to describe the long-time asymptotic behavior of the solution. However, for the present time, asymptotic formulas as well as their justifications seem to be not sufficiently clear and rigorous. Most recently an implementation of the rigorous RH scheme to the focusing nonlinear Schrödinger equation with non self-adjoint Lax operator was presented in [47–49]. In the short note [40], the initial-value problem qt + 6q2qx + qxxx = 0 (1.1) q(x, 0) = q0(x) → { cr, x → +∞, cl, x → −∞ (1.2) was considered. In [40], the solution of the problem is described by a modulated two-gap solution of the mKdV equation that corresponds to the long-time dy- namics of the compression wave when −6c2 l t + 12c2 rt < x < 4c2 l t + 2c2 rt that has not been proved up to now. The goal of this paper is to justify this statement in a transparent form by using a suitable matrix Riemann–Hilbert problem and corresponding steepest descent method. The central point of the paper is to describe in an explicit form the so-called g-function mechanism which allows to deform the original oscillatory matrix Riemann–Hilbert problem to the solvable model forms. We emphasize that our formula for a hyper-elliptic wave is written in an explicit form via theta functions. 2. Jost Solutions of Lax Equations To study the initial value problem (1.1)–(1.2) we use the Lax representation of the mKdV equation [2, 3] in the form of the over-determined system of differential equations Φx + ikσ3Φ = Q(x, t)Φ, (2.1) Φt + 4ik3σ3Φ = Q̂(x, t, k)Φ, (2.2) Journal of Mathematical Physics, Analysis, Geometry, 2012, vol. 8, No. 1 39 V. Kotlyarov and A. Minakov where Φ = Φ(x, t, k) is a 2× 2 matrix-valued function, σ3 := ( 1 0 0 −1 ) , Q(x, t) := ( 0 q(x, t) −q(x, t) 0 ) , Q̂(x, t, k) = 4k2Q(x, t, k)− 2ik(Q2(x, t, k) + Qx(x, t, k))σ3 + 2Q3(x, t, k)−Qxx(x, t, k), and k ∈ C. Equations (2.1) and (2.2) are compatible if and only if the func- tion q(x, t) satisfies the mKdV equation (1.1). To apply the inverse scattering transform to the problem (1.1)–(1.2) we have to define the matrix valued Jost solutions of the Lax equations. We define them as the solutions of the compatible equations (2.1) and (2.2) satisfying the asymptotic conditions Φr(x, t, k) = Er(x, t, k) + o (1) , x → +∞, Im k = 0, (2.3) Φl(x, t, k) = El(x, t, k) + o (1) , x → −∞, Im k = 0. (2.4) Here El(x, t, k), Er(x, t, k) are the solutions of the linear differential equations Ex + ikσ3E = QcE, Et + 4ik3σ3E = Q̂c(k)E, where c = cl and c = cr, respectively, and the constant matrix coefficients Qc and Q̂c(k) are as follows: Qc := ( 0 c −c 0 ) , Q̂c(k) = 4k2Qc − 2ikQ2 cσ3 + 2Q3 c . We choose the solutions El(x, t, k), Er(x, t, k) in the form El,r(x, t, k) = 1 2   κl,r(k) + 1 κl,r(k) κl,r(k)− 1 κl,r(k) κl,r(k)− 1 κl,r(k) κl,r(k) + 1 κl,r(k)   e−ixXl,r(k)σ3−itΩl,r(k)σ3 , where Xl,r(k) = √ k2 + c2 l,r , Ωl,r(k) = 2(2k2 − c2 l,r)Xl,r(k) , κl,r(k) = 4 √ k − icl,r k + icl,r . (2.5) The branches of the roots are fixed by the conditions Xl,r(1) > 0, κl,r(∞) = 1. Then the functions Xl,r(k) and κl,r(k) are analytic in C\[ic,−ic], where c = cl or c = cr, respectively. 40 Journal of Mathematical Physics, Analysis, Geometry, 2012, vol. 8, No. 1 Step-Initial Function to the mKdV Equation Solutions (2.3), (2.4) can be represented in the forms Φl(x, t, k) = El(x, t, k) + x∫ −∞ Kl(x, y, t)El(y, t, k)dy, Im k = 0, (2.6) Φr(x, t, k) = Er(x, t, k) + ∞∫ x Kr(x, y, t)Er(y, t, k)dy, Im k = 0, (2.7) where the kernels Kl,r(x, y, t) are sufficiently smooth and decrease to zero rapidly as x + y → ±∞. Omitting the details of the proof of these representations, we formulate below the properties of the solutions. The matrices Φl(x, t, k) and Φr(x, t, k), defined by (2.6), (2.7) and their columns Φlj(x, t, k) and Φrj(x, t, k), j = 1, 2, have the following properties: 1) determinants are equal to the identity matrix: detΦl,r(x, t, k) = 1; 2) analyticity: Φr1(x, t, k) is analytic in k ∈ Dr− := C− \ [0,−icr], Φr2(x, t, k) is analytic in k ∈ Dr+ := C+ \ [0, icr], Φl1(x, t, k) is analytic in k ∈ Dl+ := C+ \ [0, icl], Φl2(x, t, k) is analytic in k ∈ Dl− := C− \ [−icl, 0]; 3) continuity: Φr1(x, t, k) is continuous for k ∈ Dr− ∪ (−icr, icr)− ∪ (−icr, icr)+, Φr2(x, t, k) is continuous for k ∈ Dr+ ∪ (−icr, icr)− ∪ (−icr, icr)+, Φl1(x, t, k) is continuous for k ∈ Dl+ ∪ (−icl, icl)− ∪ (−icl, icl)+, Φl2(x, t, k) is continuous for k ∈ Dl− ∪ (−icl, icl)− ∪ (−icl, icl)+, where (−icl,r, icl,r)− and (−icl,r, icl,r)+ are the left- and the right-hand sides of the interval (−icl,r, icl,r); 4) symmetries: Φ22(x, t, k) = Φ11(x, t, k), Φ22(x, t,−k) = Φ11(x, t, k), Φ12(x, t, k) = −Φ21(x, t, k), Φ12(x, t,−k) = −Φ21(x, t, k), Φjl(x, t,−k) = Φjl(x, t, k), j, l = 1, 2, where Φ(x, t, k) denotes Φl(x, t, k) or Φr(x, t, k); 5) large k asymptotics: Φr1(x, t, k)e+ikx+4ik3t Φl2(x, t, k)e−ikx−4ik3t } = 1 + O ( 1 k ) , k →∞, Im k ≤ 0, Journal of Mathematical Physics, Analysis, Geometry, 2012, vol. 8, No. 1 41 V. Kotlyarov and A. Minakov Φl1(x, t, k)e+ikx+4ik3t Φr2(x, t, k)e−ikx−4ik3t } = 1 + O ( 1 k ) , k →∞, Im k ≥ 0; 6) jump: Φ−(x, t, k) = Φ+(x, t, k) ( 0 i i 0 ) , k ∈ (ic,−ic), where Φ(x, t, k) and c denote Φl(x, t, k) and cl or Φr(x, t, k) and cr, respec- tively, and Φ±(x, t, k) are the non-tangential boundary values of matrix Φ(x, t, k) from the left (−) and from the right (+) of the downward-oriented interval (−ic, ic). The matrices Φl(x, t, k) and Φr(x, t, k) are solutions of equations (2.1) and (2.2). Hence they are linear dependent, i.e.,there exists the independent of x, t matrix T (k) = Φ−1 r (x, t, k)Φl(x, t, k), k ∈ R, (2.8) which is defined for those k for which ImXr(k) = 0. Some elements of this matrix have an extended domain of definition. Indeed, using (2.8), we can find T11(k) = det(Φl1, Φr2), T21(k) = det(Φr1,Φl1), T12(k) = det(Φl2, Φr2), T22(k) = det(Φr1,Φl2). Then the above properties of the solutions Φr(x, t, k) and Φl(x, t, k) imply: • T11(k) is analytic in k ∈ C+\[0, icl] and has a continuous extension to (0, icl)− ⋃ (0, icl)+; • T22(k) is analytic in k ∈ C−\[0, icl] and has a continuous extension to (−icl, 0)− ⋃ (−icl, 0)+; • T21(k) is continuous in k ∈ (−∞, 0) ⋃ (0,−icl)− ⋃ (−icl, 0)+ ⋃ (0, +∞); • T12(k) is continuous in k ∈ (−∞, 0) ⋃ (0, icl)− ⋃ (icl, 0)+ ⋃ (0, +∞), where, as before, the signs − and + denote the left- and the right-hand sides of the intervals; • T22(k) = T11(k), T22(−k) = T11(k), • T12(k) = −T21(k), T12(−k) = −T21(k), • Tjk(−k) = Tjk(k), j, k = 1, 2. 42 Journal of Mathematical Physics, Analysis, Geometry, 2012, vol. 8, No. 1 Step-Initial Function to the mKdV Equation Denote a(k) = T11(k), b(k) = T21(k). Define the reflection coefficient r(k) = b(k) a(k) . It has the property r(−k) = r(k). The columns of the matrices Φl and Φr satisfy the following jump conditions: 7) (Φl1)−(x, t, k) a−(k) − (Φl1)+(x, t, k) a+(k) = f1(k)Φr2(x, t, k), k ∈ (icr, icl); 8) (Φl2)−(x, t, k) a−(k) − (Φl2)+(x, t, k) a+(k) = f2(k)Φr1(x, t, k), k ∈ (−icr,−icl), where f1(k) = i a−(k)a+(k) , k ∈ (0, icl), f2(k) = −f1(k), k ∈ (−icl, 0). 3. The Basic Riemann–Hilbert Problem The scattering relation (2.8) between the matrix-valued functions Φl(x, t, k) and Φr(x, t, k) and jump conditions 6, 7, 8 can be rewritten in terms of the Riemann–Hilbert problem. To do this, define the matrix-valued function M(ξ, t, k) =    ( Φl1(x, t, k) a(k) eitθ(k,ξ), Φr2(x, t, k)e−itθ(k,ξ) ) , k ∈ C+\[0, icl], ( Φr1(x, t, k)eitθ(k,ξ), Φl2(x, t, k) a(k) e−itθ(k,ξ) ) , k ∈ C−\[−icl, 0], (3.1) where x = 12ξt and θ(k, ξ) = 4k3 + 12kξ (ξ = x/12t). Below we restrict our consideration to the simplest shock problem where the initial function is discon- tinuous and piece-wise constant (pure step function): q0(x) = { cr, x ≥ 0, cl, x < 0. (3.2) Journal of Mathematical Physics, Analysis, Geometry, 2012, vol. 8, No. 1 43 V. Kotlyarov and A. Minakov Then a(k) = 1 2 ( κ(k) + 1 κ(k) ) , b(k) = 1 2 ( κ(k)− 1 κ(k) ) , r(k) = κ2(k)− 1 κ2(k) + 1 (3.3) are analytic in k ∈ C\ ([−icl,−icr] ∪ [icl, icr]), since the function κ(k) := κl(k) κr(k) (see (2.5)) is analytic in this domain. The transition coefficient a−1(k) is bounded in k ∈ C+\[icl, icr] because the function a(k) equals zero nowhere, and hence the set of eigenvalues of the linear problem (2.1) is empty. We have that f1(k) ≡ f2(k), and so we define f(k) := f1(k) = f2(k). We also have f(k) = r−(k)− r+(k), k ∈ (−icl,−icr) ∪ (icr, icl). (3.4) k r ic r ic- l ic- l ic Fig. 1. Oriented contour Σ. Let us define the oriented contour Σ = R ∪ (icl,−icl) as in Fig. 1. Then the matrix (3.1) solves the next Riemann–Hilbert problem: • the matrix-valued function M(ξ, t, k) is analytic in the domain C \ Σ; • M(ξ, t, k) is bounded in the neighborhood of the branching points icl, icr, −icl, −icr and at the origin (k = 0); • M−(ξ, t, k) = M+(ξ, t, k)J(ξ, t, k), k ∈ Σ \ {0}, 44 Journal of Mathematical Physics, Analysis, Geometry, 2012, vol. 8, No. 1 Step-Initial Function to the mKdV Equation where J(ξ, t, k) = ( 1 r(k)e−2itθ(k,ξ) −r(k)e2itθ(k,ξ) 1 + |r(k)|2 ) , k ∈ R\{0}, (3.5) = ( 1 0 f(k)e2itθ(k,ξ) 1 ) , k ∈ (icr, icl), (3.6) = ( 1 f(k)e−2itθ(k,ξ) 0 1 ) , k ∈ (−icr,−icl), (3.7) = ( ir(k) ie−2itθ(k,ξ) f(k)e2itθ(k,ξ) −ir(k) ) , k ∈ (0, icr), (3.8) = ( −ir(k) f(k)e−2itθ(k,ξ) ie2itθ(k,ξ) ir(k) ) , k ∈ (0,−icr); (3.9) • M(ξ, t, k) = I + O(k−1), k →∞, where r(k) = −r(k̄) = −r(−k) is given in (3.3), and f(k) in (3.4). If the initial function is arbitrary step-like, then a(k) may have zeroes in the domain of analyticity. In this case the matrix M(ξ, t, k) is meromorphic and residue relations between the columns of the matrix M(ξ, t, k) must be added. In what follows we suppose that the solution q(x, t) of the shock problem (1.1)–(1.2) with the pure step initial function (3.2) does exist. The above Rie- mann–Hilbert problem gives q(x, t) in the form q(x, t) = 2i lim k→∞ k[M(x/12t, t, k)]12, (3.10) where [M(x/12t, t, k)]12 is the appropriate entry of the matrix M(x/12t, t, k). 4. Long-Time Asymptotic Analysis of the Riemann–Hilbert Problem The jump matrices J(ξ, t, k) in (3.5)–(3.9) depend on exp{±2itθ(k, ξ)}. The phase function θ(k, ξ) and the signature table of its imaginary part play a very important role. For a vanishing initial function the phase function θ(k, ξ) allows to use successfully the steepest descent method for oscillatory RH problem [5] when the conjugation contour Σ coincides with the real axis R. For a non- vanishing initial function, the phase function θ(k, ξ) does not allow to carry out the asymptotic analysis of the RH problem because the contour Σ contains the segment [icl,−icl] which imposes extra (bad) properties of the phase function (indeed, e2itθ(k,ξ) grows exponentially). Therefore, we have to change the phase Journal of Mathematical Physics, Analysis, Geometry, 2012, vol. 8, No. 1 45 V. Kotlyarov and A. Minakov function θ(k, ξ) with a new one. In what follows we will use the phase function g(k, ξ) which takes different forms in different regions. A. Construction of the phase function 1. Regions ξ < −c2 l 2 + c2 r and ξ > c2 l 3 + c2 r 6 . The asymptotic analysis used for studying the asymptotic behavior in these regions is similar to those given in [23, 25, 26]. Therefore we only mention that the suitable phase functions are given by the formulas g(k, ξ) = 12ξXcl (k) + 2(2k2 − c2 l )Xcl (k), ξ > c2 l 3 + c2 r 6 , g(k, ξ) = 12ξXcr(k) + 2(2k2 − c2 r)Xcr(k), ξ < −c2 l 2 + c2 r , where Xc(k) = √ k2 + c2 is holomorphic outside the segment [ic,−ic]. We obtain the following Theorem 4.1. For t → ∞ and x > (4c2 l + 2c2 r)t the solution of the problem (1.1)–(1.2) with the initial pure step function (3.2) takes the form q(x, t) = cr + O(e−Ct), where C > 0 is some positive constant. Theorem 4.2. For t →∞ and x < (−6c2 l +12c2 r)t the solution of the problem (1.1)–(1.2) with the initial pure step function (3.2) takes the form q(x, t) = cl + O(t−1/2). In what follows we will deal only with the 2. Region −c2 l 2 + c2 r < ξ < c2 l 3 + c2 r 6 ( (−6c2 l +12c2 r)t < x < (4c2 l +2c2 r)t ) . In this region we use the function g(k, ξ) with the following properties: (1) g(k, ξ) is analytic in the domain k ∈ C\[icl,−icl]; (2) ∃ lim k→∞ (g(k, ξ)− θ(k, ξ)) = g0(ξ) ∈ C; (3) the set {k : Im g(k, ξ) = 0} divides the complex plane into four connected open sets and contains necessarily the set R ∪ [icl, id] ∪ [icr,−icr] ∪ [−id,−icl], where d = d(ξ) ∈ [icl, icr] is some function of ξ. We look for such a function in the form g(k, ξ) = k∫ icl 12k(k2 + µ2)(k2 + d2)dk w(k, ξ) , w(k, ξ) = √ (k2 + c2 l )(k 2 + d2)(k2 + c2 r) , 46 Journal of Mathematical Physics, Analysis, Geometry, 2012, vol. 8, No. 1 Step-Initial Function to the mKdV Equation where function w(k, ξ) is positive on the positive part of the real axis and analytic in k ∈ C\ ([icl, id] ∪ [icr,−icr] ∪ [−id,−icl]). Here unknown numbers d and µ have to be determined as the functions of ξ. The integration contour is chosen to have no intersection with the segment [icl,−icl]. It is easy to see that g(k, ξ) ∈ R if k lies on the left- or right-hand side of the segment [icl, id]. To satisfy the requirement g(k, ξ) ∈ R, if k lies on the left- or right-hand side of [icr,−icr] ∪ [−id,−icl], we have to choose such numbers µ and d that ∫ id icr dg(k, ξ) = 0 and ∫ −icr −id dg(k, ξ) = 0. Due to the symmetry of dg(k, ξ) under the change of variable k 7→ −k, we can see that the last two requirements are equivalent to each other and can be written as follows: d∫ cr y(y2 − µ2) √ d2 − y2dy√ (c2 l − y2)(y2 − c2 r) = 0. (4.1) This formula defines µ = µ(d) as a strictly increasing function on the segment [cr, cl], and µ(cr) = cr , µ(cl) = √ c2 l + 2c2 r 3 . (4.2) Indeed, by expressing µ in d through formula (4.1) and then by taking the first derivative of µ2(d), we find ( µ2 )′ d = d d∫ cr ρ(y)h1(y)dy d∫ cr ρ(y)h2(y)dy − d∫ cr ρ(y)h1(y)h2(y)dy d∫ cr ρ(y)dy   d∫ cr h1(y)h2(y)ρ(y)dy   2 , where we denote ρ(y) = y√ (c2 l − y2)(d2 − y2)(y2 − c2 r) , h1(y) = y2, h2(y) = d2 − y2 . Let us note that all the three functions ρ(.), h1(.), h2(.) are positive on the segment [cr, d], moreover, h1(.) is increasing on the segment and h2(.) is decreasing on the segment. Therefore we can use Journal of Mathematical Physics, Analysis, Geometry, 2012, vol. 8, No. 1 47 V. Kotlyarov and A. Minakov Lemma 4.1. Let ρ(.), h1(.), h2(.) be positive functions on the segment [a, b] ⊂ R such that the integrals of their combinations ρ(y), ρ(y)h1(y), ρ(y)h2(y), ρ(y)h1(y)h2(y) are convergent in proper or improper sense. Let also assume that h1 is the increasing function on the segment, and h2 is decreasing function. Then d∫ cr ρ(y)h1(y)dy d∫ cr ρ(y)h2(y)dy − d∫ cr ρ(y)h1(y)h2(y)dy d∫ cr ρ(y)dy ≥ 0. We will prove this lemma in the Appendix. Now we want to satisfy the requirement (2) lim k→∞ (g(k, ξ)− θ(k, ξ)) ∈ C, which is fulfilled if dg(k, ξ)−dθ(k, ξ) = O(k−2)dk, as k →∞, is fulfilled. Since dθ(k, ξ) = 12(k2 + ξ)dk and dg(k, ξ) = [ 12k2 − 6(c2 l + c2 r − d2 − 2µ2) + O(k−2) ] dk as k → ∞, we need ξ + c2 l + c2 r 2 = µ2 + d2 2 . (4.3) Equations (4.2) yield that µ2(d) + d2 2 − c2 l + c2 r 2 varies over the segment [ −c2 l 2 + c2 r, c2 l 3 + c2 r 6 ] when d varies over the segment [cr, cl]. So, from (4.3) we get that for any ξ ∈ [ −c2 l 2 + c2 r, c2 l 3 + c2 r 6 ] there exists a single d = d(ξ) ∈ [cr, cl] such that (4.1) and (4.3) are fulfilled. Equality (4.3) implies that d = d(ξ) is a continuous function. Thus, the function g(k, ξ) is completely defined and it has the property (2a) lim k→∞ (g(k, ξ)− θ(k, ξ)) = 0 which follows from the existence of the limit and the relations below: g(k, ξ)− θ(k, ξ) ∈ iR , k ∈ (icl,+i∞) ; g(k, ξ)− θ(k, ξ) ∈ R , k ∈ R . Besides, (4) g−(k, ξ) + g+(k, ξ) = 0 , k ∈ (icl, id) ∪ (icr,−icr) ∪ (−id,−icl); (5) g−(k, ξ)− g+(k, ξ) = Bg(ξ) , k ∈ (id, icr) ∪ (−icr,−id) , where Bg(ξ) = 2 icl∫ id dg+(k, ξ) = 2 −icl∫ −id dg+(k, ξ) > 0. (4.4) The signature table of the imaginary part of the function g(k, ξ) is given in Fig. 2. 48 Journal of Mathematical Physics, Analysis, Geometry, 2012, vol. 8, No. 1 Step-Initial Function to the mKdV Equation Fig. 2. The signature table of Im g(k, ξ). 3. Changing of the phase function. As the phase function θ(k, ξ) is not suitable now, the Riemann–Hilbert problem for the matrix M(ξ, t, k) has to be considered with a new phase function g(k, ξ). Let us define the new matrix- function M (1)(ξ, t, k) = M(ξ, t, k)G(1)(ξ, t, k), where G(1)(ξ, t, k) = eit(g(k,ξ)−θ(k,ξ))σ3 . Then the function M (1)(ξ, t, k) solves the RH problem M (1) − (ξ, t, k) = M (1) + (ξ, t, k)J (1)(ξ, t, k), k ∈ Σ1 := Σ, M (1)(ξ, t, k) → I, k →∞, where J (1)(ξ, t, k) = ( 1 r(k)e−2itg(k,ξ) −r(k)e2itg(k,ξ) 1 + |r(k)|2 ) , k ∈ R\{0}, (4.5) = ( eit(g−(k,ξ)−g+(k,ξ)) 0 f(k)eit(g−(k,ξ)+g+(k,ξ)) e−it(g−(k,ξ)−g+(k,ξ)) ) , k ∈ (icr, icl), (4.6) Journal of Mathematical Physics, Analysis, Geometry, 2012, vol. 8, No. 1 49 V. Kotlyarov and A. Minakov = ( eit(g−(k,ξ)−g+(k,ξ)) f(k)e−it(g−(k,ξ)+g+(k,ξ)) 0 e−it(g−(k,ξ)−g+(k,ξ)) ) , k ∈ (−icr,−icl), (4.7) = ( ir(k)eit(g−(k,ξ)−g+(k,ξ)) ie−it(g−(k,ξ)+g+(k,ξ)) f(k)eit(g−(k,ξ)+g+(k,ξ)) − r(k)e−it(g−(k,ξ)−g+(k,ξ)) ) , k ∈ (0, icr), (4.8) = ( −ir(k)eit(g−(k,ξ)−g+(k,ξ)) f(k)e−it(g−(k,ξ)+g+(k,ξ)) ieit(g−(k,ξ)+g+(k,ξ)) ir(k)e−it(g−(k,ξ)−g+(k,ξ)) ) , k ∈ (0,−icr). (4.9) 4. Transferring of the jump contour from the real line. Define a decom- position of the k-complex plane into domains Ωj , j = 1, 2, 3, 4 as shown in Fig. 3. l ic l ic- r ic- r ic id id- 1 L 1 L 2 L 2 L Fig. 3. The contour Σ2. Here the contour L1 lies in the part of the complex plane, where Img(k, ξ) > 0, and L2 lies in the part of the complex plane, where Img(k, ξ) < 0. The transformation below transfers the jump contour from the real line M (2)(ξ, t, k) = M (1)(ξ, t, k)G(2)(ξ, t, k), where G(2)(ξ, t, k) = ( 1 0 −r(k)e2itg(k,ξ) 1 ) , k ∈ Ω1 , (4.10) = ( 1 −r(k)e−2itg(k,ξ) 0 1 ) , k ∈ Ω2 , (4.11) = I, k ∈ (Ω1 ∪ Ω2) C . (4.12) 50 Journal of Mathematical Physics, Analysis, Geometry, 2012, vol. 8, No. 1 Step-Initial Function to the mKdV Equation G(2) — transformation leads to the RH-problem M (2) − (ξ, t, k) = M (2) + (ξ, t, k)J (2)(ξ, t, k) , k ∈ Σ2, M (2)(ξ, t, k) → I , k →∞ , where J (2)(ξ, t, k) = ( G (2) + )−1 J (1)(ξ, t, k)G(2) − (ξ, t, k). Taking into account the definition of G(2) (4.10)–(4.12), J (1) (4.5)–(4.9) and the property a2(k)− b2(k) = 1, we get J (2)(ξ, t, k) = J (1)(ξ, t, k), k ∈ (icl, id) ∪ (−id,−icl), = eit(g−(k,ξ)−g+(k,ξ))σ3 = eitBg(ξ)σ3 , k ∈ (icr, id) ∪ (−id,−icr), = ( 0 i i 0 ) eit(g−(k,ξ)+g+(k,ξ))σ3 = ( 0 i i 0 ) , k ∈ (−icr, icr), = G(2)(ξ, t, k), k ∈ L1, = ( G(2) )−1 (ξ, t, k), k ∈ L2. 5. Next transformation. The function f(k) has the analytic continuation f̂(k) := 1 a(k)b(k) from the intervals (icl, icr)∪(−icr,−icl). Thus we can factorize the jump matrix J (2)(ξ, t, k) on the intervals (icl, id) ∪ (−id,−icl) as follows: J (2)(ξ, t, k) = F−σ3 + (k, ξ)  1 F 2 +(k, ξ)e−2itg+(k,ξ) f̂+(k) 0 1   ( 0 i i 0 ) ×  1 −F 2−(k, ξ)e−2itg−(k,ξ) f̂−(k) 0 1  F σ3− (k, ξ) , k ∈ (icl, id), (4.13) = F−σ3 + (k, ξ)   1 0 e2itg+(k,ξ) F 2 +(k, ξ)f̂+(k) 1   ( 0 i i 0 ) ×   1 0 −e2itg−(k,ξ) F 2−(k, ξ)f̂−(k) 1  F σ3− (k, ξ) , k ∈ (−icl,−id). (4.14) Journal of Mathematical Physics, Analysis, Geometry, 2012, vol. 8, No. 1 51 V. Kotlyarov and A. Minakov Direct calculations show that the above is possible if: • F (k, ξ) is analytic outside the segment [icl,−icl]; • F (k, ξ) does not vanish in the complex plane with the cut along [icl,−icl]; • F (k, ξ) satisfies the jump relations F+(k, ξ)F−(k, ξ) =    −if(k), k ∈ (icl, id), i f(k) , k ∈ (−id,−icl), 1 , k ∈ (icr,−icr), F+(k, ξ) = F−(k, ξ)ei∆(ξ) , k ∈ (id, icr) ∪ (−icr,−id), where ∆(ξ) is some function of ξ, which has to be determined; • F (k, ξ) is bounded at the infinity; • F (k, ξ)a(k) is bounded in a small neighborhood of the point icl; • F (k, ξ)a−1(k) is bounded in a small neighborhood of the point −icl; • F (k, ξ) is bounded in small neighborhoods of the points ±id, ±icr, 0. To solve this conjugation problem we use the function w(k) = √ (k2 + c2 l )(k 2 + d2)(k2 + c2 r). Let us note that −if(k) = 1 a−(k)a+(k) > 0. The jump relations on F can be rewritten in the form: [ log F (k, ξ) w(k, ξ) ] + − [ log F (k, ξ) w(k, ξ) ] − = − log (a−(k)a+(k)) w+(k, ξ) , k ∈ (icl, id) , [ log F (k, ξ) w(k, ξ) ] + − [ log F (k, ξ) w(k, ξ) ] − = log (a−(k)a+(k)) w+(k, ξ) , k ∈ (−id,−icl) , [ log F (k, ξ) w(k, ξ) ] + − [ log F (k, ξ) w(k, ξ) ] − = i∆(ξ) w(k, ξ) , k ∈ (id, icr) ∪ (−icr,−id). The function F (k, ξ) =exp    w(k) 2πi id∫ icl − log (a+(s)a−(s)ds) (s− k)w+(s)    exp    w(k) 2πi −icl∫ −id log (a+(s)a−(s)ds) (s− k)w+(s)    × exp    i∆(ξ)w(k) 2πi icr∫ id ds (s− k)w(s)    exp    i∆(ξ)w(k) 2πi −id∫ −icr ds (s− k)w(s)    52 Journal of Mathematical Physics, Analysis, Geometry, 2012, vol. 8, No. 1 Step-Initial Function to the mKdV Equation satisfies the first, the second and the third properties. To make F (k, ξ) bounded at the infinity we have to expand F (k, ξ) in series when k → ∞. Since for the integer n id∫ icl −sn log (a+(s)a−(s)) ds w+(s) = (−1)n+1 −icl∫ −id sn log (a+(s)a−(s)) ds w+(s) , icr∫ id snds w(s) = (−1)n+1 −id∫ −icr snds w(s) , then the behavior of log F (k, ξ) at the infinity is described by the asymptotic formula log F (k, ξ) = 2k   1 2πi id∫ icl s log (a+(s)a−(s)) ds w+(s, ξ) − −i∆(ξ) 2πi icr∫ id sds w(s, ξ)   + o(1). So, we put ∆(ξ) = −i id∫ icl s log (a+(s)a−(s)ds) w+(s, ξ)   icr∫ id sds w(s, ξ)   −1 . (4.15) It is easy to check that ∆(ξ) ∈ R. Thus the function F (k, ξ) satisfies all the requirements. Using factorizations (4.13) and (4.14) and the transformation M (3)(ξ, t, k) = M (2)(ξ, t, k)G(3)(ξ, t, k) , where G(3)(ξ, t, k) = F−σ3(k, ξ)  1 F 2(k, ξ)e−2itg(k,ξ) f̂(k) 0 1   , k ∈ Ω5 ∪ Ω7 , = F−σ3(k, ξ)   1 0 e2itg(k,ξ) F 2(k, ξ)f̂(k) 1   , k ∈ Ω6 ∪ Ω8 , = F−σ3 , k /∈ (Ω5 ∪ Ω6 ∪ Ω7 ∪ Ω8) , we obtain the RH problem M (3) − (ξ, t, k) = M (3) + (ξ, t, k)J (3)(ξ, t, k), k ∈ Σ3, M (3)(ξ, t, k) → I , k →∞. Journal of Mathematical Physics, Analysis, Geometry, 2012, vol. 8, No. 1 53 V. Kotlyarov and A. Minakov l ic l ic- r ic- r ic id id- 1 L 1 L 2 L 2 L 8 L 6 L 5 L 7 L Fig. 4. The contour Σ3. The jump matrix J (3)(ξ, t, k) = ( G (3) + )−1 J (2)(ξ, t, k)G(3)(ξ, t, k) is J (3)(ξ, t, k) = ( 1 0 −r(k)F−2(k, ξ)e2itg(k,ξ) 1 ) , k ∈ L1 , = ( 1 r(k)F 2(k, ξ)e−2itg(k,ξ) 0 1 ) , k ∈ L2 , J (3)(ξ, t, k) =  1 F 2(k, ξ)e−2itg(k,ξ) f̂(k) 0 1   , k ∈ L7 , =  1 −F 2(k, ξ)e−2itg(k,ξ) f̂(k) 0 1   , k ∈ L5 , J (3)(ξ, t, k) =   1 0 e2itg(k,ξ) F 2(k, ξ)f̂(k) 1   , k ∈ L8 , =   1 0 −e2itg(k,ξ) F 2(k, ξ)f̂(k) 1   , k ∈ L6 , 54 Journal of Mathematical Physics, Analysis, Geometry, 2012, vol. 8, No. 1 Step-Initial Function to the mKdV Equation J (3)(ξ, t, k) = e(itBg(ξ)+i∆(ξ))σ3 , k ∈ (id, icr) ∪ (−icr,−id) , = ( 0 i i 0 ) , k ∈ (icl, id) ∪ (icr,−icr) ∪ (−id,−icl). 6. Model problem. Now we consider a model problem M (mod) − (ξ, t, k) = M (mod) + (ξ, t, k)J (mod)(ξ, t, k), M (mod)(ξ, t, k) → I as k →∞, where J (mod)(ξ, t, k) =    ( eitBg(ξ)+i∆(ξ) 0 0 e−itBg(ξ)−i∆(ξ) ) , k ∈ (id, icr) ∪ (−icr,−id) , ( 0 i i 0 ) , k ∈ (icl, id) ∪ (icr,−icr) ∪ (−id,−icl). (4.16) To solve the model problem (4.16), we introduce the Riemann surface X, which is given by w2(k) = (k2 + c2 l )(k 2 + d2)(k2 + c2 r). We will use a realization of this algebraic curve as the two-sheet Riemann surface. The upper and lower sheets of the surface are two complex planes merged along the cuts [icl, id], [icr,−icr] and [−id,−icl]. On the upper sheet of this surface w(1) > 0. The basis {a1, b1, a2, b2} of cycles of this Riemann surface is as follows. The a1-cycle starts from the right-hand side of the cut [icl, id] on the upper sheet, goes to the right-hand side of the cut [icr,−icr], proceeds to the lower sheet and then returns to the starting point. The b1-cycle is a closed counter clock-wise oriented simple loop around the cut [icl, id]. The a2-cycle starts from the right- hand side of the cut [id,−id], [icr,−icr] on the upper sheet, goes to the right-hand side of the cut [−id,−icl], proceeds to the lower sheet and then returns to the starting point. The b2-cycle is a closed counter clock-wise oriented simple loop around the segment [icl,−icr]. The basis dω = ( dω1 dω2 ) of the normalized holomorphic differentials on X has the form dω1 = πi kdk w(k)   ∫ a1 kdk w(k)   −1 + πi dk w(k)   ∫ a1 dk w(k)   −1 , Journal of Mathematical Physics, Analysis, Geometry, 2012, vol. 8, No. 1 55 V. Kotlyarov and A. Minakov dω2 = πi kdk w(k)   ∫ a1 kdk w(k)   −1 − πi dk w(k)   ∫ a1 dk w(k)   −1 . Then ∫ aj dωl = 2πiδjl, B = B(ξ) =    Bjl = ∫ bj dωl     = ( B1 B2 B2 B1 ) , where B1 := ∫ b1 dω1 , B2 := ∫ b1 dω2 , and B1 < B2 < 0 , and theta function Θ(z) = Θ(z|B(ξ)) = ∑ m∈Z2 exp { 1 2 (B(ξ)m, m) + (z, m) } , z ∈ C2, has the property Θ(z + 2πin + B(ξ)l) = Θ(z) exp { −1 2 (B(ξ)l, l)− (z, l) } , n ∈ Z2, l ∈ Z2. Now we introduce the Abel map on X A : X → C2/ ( 2πiZ2 + B(ξ)Z2 ) , A(P ) = P∫ icl dω (4.17) and the functions ϕ(k, ξ), ψ(k, ξ) : {the first sheet of the X} → C ϕj(k, ξ) = Θ(A(k)−A(Dj)−K − (itBg(ξ) + i∆(ξ)) (1, 1)T ) Θ(A(k)−A(Dj)−K) , ψj(k, ξ) = Θ(−A(k)−A(Dj)−K − (itBg(ξ) + i∆(ξ)) (1, 1)T ) Θ(−A(k)−A(Dj)−K) , j = 1, 2. (4.18) Here D1 = P1 + P2 is the divisor consisting of two points on the lower sheet, P1 = i √ clcrd cl + cr − d and P2 = −i √ clcrd cl + cr − d , D2 = τD1 lies on the upper sheet, and A(D1) = −A(D2). The vector K is the Riemann constant of the surface X, Bg(ξ) and ∆(ξ) are defined in (4.4) and 56 Journal of Mathematical Physics, Analysis, Geometry, 2012, vol. 8, No. 1 Step-Initial Function to the mKdV Equation (4.15). The integration contour in (4.17) is taken from the upper sheet and it does not intersect the interval (−icl, icl). The functions (4.18) have the following properties: ϕj+(k, ξ) = ψj−(k, ξ), ψj+(k, ξ) = ϕj−(k, ξ), k ∈ (icl, id) ∪ (icr,−icr) ∪ (−id,−icl). ϕj−(k, ξ) = ϕj+(k, ξ)eitBg(ξ)+i∆(ξ), ψj−(k, ξ) = ψj+(k, ξ)e−itBg(ξ)−i∆(ξ), k ∈ (id, icr) ∪ (−icr,−id). Define a function γ(k) = γ(k, ξ) = 4 √ k − icl k − id 4 √ k − icr k + icr 4 √ k + id k + icl , which is analytic outside the union of segments [icl, id] ∪ [icr,−icr] ∪ [−id,−icl] and satisfies the jump conditions γ−(k, ξ) = iγ+(k, ξ), k ∈ [icl, id] ∪ [icr,−icr] ∪ [−id,−icl]. Then the solution of the model problem (4.16) can be written as follows: M (mod)(ξ, t, k) =   M (mod) 11 (ξ, t, k) M (mod) 12 (ξ, t, k) M (mod) 21 (ξ, t, k) M (mod) 22 (ξ, t, k)   , M (mod) 11 (ξ, t, k) = 1 2 ( γ(k, ξ) + 1 γ(k, ξ) ) ϕ1(k, ξ) ϕ1(∞, ξ) , M (mod) 12 (ξ, t, k) = 1 2 ( γ(k, ξ)− 1 γ(k, ξ) ) ψ1(k, ξ) ϕ1(∞, ξ) , M (mod) 21 (ξ, t, k) = 1 2 ( γ(k, ξ)− 1 γ(k, ξ) ) ϕ2(k, ξ) ψ2(∞, ξ) , M (mod) 22 (ξ, t, k) = 1 2 ( γ(k, ξ) + 1 γ(k, ξ) ) ψ2(k, ξ) ψ2(∞, ξ) . Then, by the formula (3.10), qmod(x, t) := lim k→∞ 2ik ( M (mod) ( x 12t , t, k ) − I ) 21 =(cl − d(ξ) + cr) Θ(A(∞) + A(D1)−K − (itBg(ξ) + i∆(ξ))(1, 1)T ) Θ(A(∞) + A(D1)−K) × Θ(−A(∞) + A(D1)−K) Θ(−A(∞) + A(D1)−K − (itBg(ξ) + i∆(ξ))(1, 1)T ) . Journal of Mathematical Physics, Analysis, Geometry, 2012, vol. 8, No. 1 57 V. Kotlyarov and A. Minakov Theorem 4.3. Let t →∞. Then in the region (12c2 r−6c2 l )t < x < (4c2 l +2c2 r)t the solution of the problem (1.1)–(1.2) with initial pure step function takes the form of a modulated hyper-elliptic wave q(x, t) = (cl − d(ξ) + cr) Θ(A(∞) + A(D1)−K − (itBg(ξ) + i∆(ξ))(1, 1)T ) Θ(A(∞) + A(D1)−K) × Θ(−A(∞) + A(D1)−K) Θ(−A(∞) + A(D1)−K − (itBg(ξ) + i∆(ξ))(1, 1)T ) + O(t−1/2). 5. Appendix Lemma. Let ρ(.), h1(.), h2(.) be positive functions on the segment [a, b] ⊂ R such that the integrals of their products ρ, ρ(y)h1(y), ρ(y)h2(y), ρ(y)h1(y)h2(y) are convergent in proper or improper sense. Let also h1 be an increasing function on the segment, and h2 be a decreasing function. Then d∫ cr ρ(y)h1(y)dy d∫ cr ρ(y)h2(y)dy − d∫ cr ρ(y)h1(y)h2(y)dy d∫ cr ρ(y)dy ≥ 0. P r o o f. If all integrals are proper, then we can approximate them by partial sums. By substituting them into the input inequality instead of integrals, we obtain the inequality 1 N2 N∑ n=1 ρ(yn)h1(yn) N∑ m=1 ρ(ym)h2(ym)− 1 N2 N∑ n=1 ρ(yn)h1(yn)h2(yn) N∑ m=1 ρ(ym) ≥ 0. Let us multiply both sides on N2 and multiply the expressions in parenthesis ∑ n, m ρ(yn)h1(yn)ρ(ym)h2(ym)− ∑ n, m ρ(yn)ρ(ym)h1(yn)h2(yn) ≥ 0. We can see that the terms of series, when n = m, disappear. Thus we obtain ∑ n<m ρ(yn)ρ(ym) (h1(yn)h2(ym) + h1(ym)h2(yn)) − ∑ n<m ρ(yn)ρ(ym) (h1(yn)h2(yn) + h1(ym)h2(ym)) ≥ 0. Then the above is equivalent to ∑ n<m ρ(yn)ρ(ym) (h1(yn)− h1(ym)) (h2(ym)− h2(yn)) ≥ 0, which is evidently true. 58 Journal of Mathematical Physics, Analysis, Geometry, 2012, vol. 8, No. 1 Step-Initial Function to the mKdV Equation If the integrals are improper, then we can approximate them by proper ones for which the lemma is proven. References [1] C.S. Gardner, J.M. Greene, M.D. Kruskal, and R.M. Miura, Korteweg–de Vries Equation and Generalization. VI. Methods for Exact Solution. — Phys. Rev. Lett. 27 (1974), 97–133. [2] S.P. Novikov, ed., Soliton Theory. Inverse Problem Method. Moskow, 1980. (Rus- sian) [3] M.J. Ablowitz and H. Segur, Solitons and the Inverse Scattering Transform. SIAM, Philadelphia, 1981. [4] P. Deift and X. Zhou, A Steepest Descent Method for Oscillatory Riemann–Hilbert Problems. — Bull. Amer. Math. Soc. (N.S.) 26 (1992), No. 1, 119–123. [5] P. Deift and X. Zhou, A Steepest Descent Method for Oscillatory Riemann–Hilbert Problems. Asymptotics for the MKdV equation. — Ann. Math. 137 (1993), No. 2, 295–368. [6] P. Deift, A. Its, and X. Zhou, Long-Time Asymptotics for Integrable Nonlinear Wave Equations. — Important Developments in Soliton Theory. Springer Ser. Non- linear Dynam., Springer, Berlin, 1993, 181–204. [7] A.S. Fokas and A.R. Its, The Linearization of the Initial-Boundary Value Problem of the Nonlinear Schrodinger Equation. — SIAM J. Math. Anal. 27 (1996), No. 3, 738–764. [8] P. Deift, A.R. Its, and X. Zhou, A Riemann–Hilbert Approach to Asymptotic Problems Arising in the Theory of Random Matrix Models, and Also in the Theory of Integrable Statistical Mechanics. — Ann. Math. 146 (1997), 149–235. [9] A.S. Fokas, A Unified Transform Method for Solving Linear and Certain Nonlinear PDEs. — Proc. Roy. Soc. London Ser. A 453 (1997), 1411–1443. [10] A.S. Fokas, A.R. Its, and L.-Y. Sung, The Nonlinear Schrodinger Equation on the Half-Line. — Nonlinearity 18 (2005), 1771–1822. [11] A.S. Fokas and C.R. Menyuk, Integrability and Self-Similarity in Transient Stimu- lated Raman Scattering. — J. Nonlinearity Sci. 9 (1999), No. 1, 1–31. [12] A. Boutet de Monvel, A.S. Fokas, and D.G. Shepelsky, The mKdV Equation on the Half-Line. — J. Inst. Math. Jussieu 3 (2004), No. 2, 139–164. [13] A. Boutet de Monvel, A.S. Fokas, and D.G. Shepelsky, Integrable Nonlinear Evo- lution Equations on a Finite Interval. — Comm. Math. Phys. 263 (2006), No. 1, 133–172. [14] A. Boutet de Monvel, and D.G. Shepelsky, Riemann–Hilbert Problem in the Inverse Scattering for the Camassa–Holm Equation on the Line. — MSRI Publications 55 (2008), 53–75. Journal of Mathematical Physics, Analysis, Geometry, 2012, vol. 8, No. 1 59 V. Kotlyarov and A. Minakov [15] A. Boutet de Monvel and D.G. Shepelsky, Long-Time Asymptotics of the Camassa– Holm Equation on the Line. — Integrable Systems and Random Matricies. Contem- porary Mathematics 458 (2008), 99–116. [16] E.A. Moscovchenko and V.P. Kotlyarov, A New Riemann–Hilbert Problem in a Model of Stimulated Raman Scattering. — J. Phys. A.: Math. Gen. 39 (2006), 14591–14610. [17] E.A. Moscovchenko, Simple Periodic Boundary Data and Riemann–Hilbert Problem for Integrable Model of the Stimulated Raman Scattering. — J. Math. Phys., Anal., Geom. 5 (2009), No. 1, 82–103. [18] A. Boutet de Monvel and D.G. Shepelsky, The Camassa–Holm Equation on the Half-Line: a Riemann–Hilbert Approach. — J. Geom. Anal. 18 (2008), No. 2, 285–323. [19] A. Boutet de Monvel, A.R. Its, and V.P. Kotlyarov, Long-Time Asymptotics for the Focusing NLS Equation with Time-Periodic Boundary Condition on the Half-Line. — Comm. Math. Phys. 290 (2009), No. 2, 479–522. [20] A. Boutet de Monvel, V.P. Kotlyarov, and D.G. Shepelsky, Decaying Long-Time Asymptotics for the Focusing NLS Equation with Periodic Boundary Condition. — Int. Math. Res. Notices 3 (2009), 547–577. [21] E.A. Moskovchenko and V.P. Kotlyarov, Periodic Boundary Data for an Integrable Model of Stimulated Raman Scattering: Long-Time Asymptotic Behavior. — J. Phys. A 43 (2010), No. 5, 31. [22] A. Boutet de Monvel, V. Kotlyarov, D.G. Shepelsky, and C. Zheng, Initial Boundary Value Problems for Integrable Systems: Towards the Long Time Asymptotics. — Nonlinearity 23 (2010), No. 10, 2483–2499. [23] V. Kotlyarov and A. Minakov, Riemann–Hilbert Problem to the Modified Korteveg– de Vries Equation: Long-Time Dynamics of the Step-Like Initial Data. — J. Math. Phys. 51 (2010), No. 9, 31 pp. [24] A. Boutet de Monvel, V.P. Kotlyarov, and D.G. Shepelsky, Focusing NLS Equation: Long-Time Dynamics of Step-Like Initial Data. — Int. Math. Res. Notices 7 (2011), 1613–1653. [25] A. Minakov, Asymptotics of Rarefaction Wave Solution to the MKdV Equation. — J. Math. Phys., Anal., Geom. 7 (2010), No. 1, 59–86. [26] A. Minakov, Long-Time Behavior of the Solution to the MKdV Equation with Step-Like Initial Data. — J. Phys. A 44 (2011), No. 8, 31 pp. [27] P. Deift, A. Its, and X. Zhou, A Riemann–Hilbert Approach to Asymptotic Prob- lems Arising in the Theory of Random Matrix Models, and also in the Theory of Integrable Statistical Mechanics. — Ann. of Math. (2) 146 (1997), No. 1, 149–235. [28] P. Deift, T. Kricherbauer, K.T.-R. McLaughlin, S. Venakides, and X. Zhou, Uni- form Asymptotics for Polynomials Orthogonal with Respect to Varying Exponential Weights and Applications to Universality Questions in Random Matrix Theory. — Commun. Pure Appl. Math. 52 (1999), No. 11, 1335–1425. 60 Journal of Mathematical Physics, Analysis, Geometry, 2012, vol. 8, No. 1 Step-Initial Function to the mKdV Equation [29] P.M. Bleher, Lectures on Random Matrix Models. The Riemann–Hilbert approach, arXiv:0801.1858 (2008). [30] A.V. Gurevich and L.P. Pitaevskii, Splitting of an Initial Jump in the KdV Equa- tion. — JETP Letters 17 (1973), No. 5, 268–271. (Russian) [31] E.Ya. Khruslov, Asymptotic Behavior of the Solution of the Cauchy Problem for the Korteweg–de Vries Equation with Steplike Initial Data. — Mat. Sb. 99 (1976), No. 2, 261–281. (Russian) [32] E.Ya. Khruslov and V.P. Kotlyarov, Solitons of the Nonlinear Schrodinger Equation, which Are Generated by the Continuous Spectrum. — Teoret. Mat. Fiz. 68 (1986), No. 2, 172–186. (Russian) [33] E.Ya. Khruslov and V.P. Kotlyarov, Soliton Asymptotics of Nondecreasing Solutions of Nonlinear Completely Integrable Evolution Equations. — Advances in Soviet Mathematics. Operator Spectral Theory and Related Topics, AMS 19 (1994), 72. [34] I.A. Anders, E.Ya. Khruslov, and V.P. Kotlyarov, Curved Asymptotic Solitons of the Kadomtsev–Petviashvili Equation. — Teoret. Mat. Fiz. 99 (1994), No. 1, 27–35. (Russian) [35] V.B. Baranetsky and V.P. Kotlyarov, Asymptotic Behavior in a Back Front Domain of the Solution of the KdV Equation with a ”Step Type” Initial Condition. — Teoret. Mat. Fiz. 126 (2001), No. 2, 214–227. (Russian) [36] A. Boutet de Monvel and V.P. Kotlyarov, Generation of Asymptotic Solitons of the Nonlinear Schrodinger Equation by Boundary Data. — J. Math. Phys. 44 (2003), No. 8, 3185–3215. [37] E.Ya. Khruslov and V.P. Kotlyarov, Generation of Asymptotic Solitons in an Inte- grable Model of Stimulated Raman Scattering by Periodic Boundary Data. — Mat. Fiz., Anal., Geom. 10 (2003), No. 3, 364–186. [38] I. Egorova, K. Grunert, and G. Teschl, On the Cauchy Problem for the Korteweg– de Vries Equation with Steplike Finite-Gap Initial Data. I. Schwartz-type Pertur- bations. — Nonlinearity 22 (2009), No. 6, 1431–1457. [39] I. Egorova, J. Michor, and G. Teschl, Inverse Scattering Transform for the Toda Hierarchy with Steplike Finite-Gap Backgrounds. — J. Math. Phys. 500 (2009), No. 10, 103521–103529. [40] R.F. Bikbaev, The Influence of Viscosity on the Structure of Shock Waves in the MKdV Model. — J. Math. Sci. 77 (1995), No. 2, 3042–3045. [41] R.F. Bikbaev, Complex Whitham Deformations in Problems with ”Integrable In- stability”. — Teoret. Mat. Fiz. 104 (1995), No. 3, 393–419. [42] R.F. Bikbaev, Modulational Instability Stabilization Via Complex Whitham De- formations: Nonlinear Schrodinger Equation. — J. Math. Sci. 85 (1997), No. 1, 1596–1604. [43] V.Yu. Novokshenov, Time Asymptotics for Soliton Equations in Problems with Step Initial Conditions. — J. Math. Sci. 125 (2005), No. 5, 717–749. Journal of Mathematical Physics, Analysis, Geometry, 2012, vol. 8, No. 1 61 V. Kotlyarov and A. Minakov [44] P.D. Lax, C.D. Levermore, and S. Venakides, The Generation and Propagation of Oscillations in Dispersive Initial Value Problems and Their Limiting Behavior. — Important Developments in Soliton Theory. Springer Ser. Nonlinear Dynam. (1993), 205–241. [45] P. Deift, S. Kamvissis, T. Kriecherbauer, and X. Zhou, The Toda Rarefaction Problem. — Comm. Pure Appl. Math. 49 (1996), No. 1, 35–83. [46] P. Deift, S. Venakides, and X. Zhou, New Results in Small Dispersion KdV by an Extension of the Steepest Descent Method for Riemann–Hilbert Problems. — IMRN 6 (1997), 286–299. [47] R. Buckingham and S. Venakides, Long-Time Asymptotics of the Nonlinear Schrodinger Equation Shock Problem. — Comm. Pure Appl. Math. 60 (1997), No. 9, 286–299. [48] A. Boutet de Monvel and V. Kotlyarov, The Focusing Nonlinear Schrodinger Equa- tion on the Quarter Plane with Time-Periodic Boundary Condition: a Riemann– Hilbert Approach. — J. Inst. Math. Jussieu 6 (2007), No. 4, 579–611. [49] A. Boutet de Monvel, A.R. Its, and V. Kotlyarov, Long-Time Asymptotics for the Focusing NLS Equation with Time-Periodic Boundary Condition. — C. R. Math. Acad. Sci. Paris 345 (2007), No. 11, 615–620. 62 Journal of Mathematical Physics, Analysis, Geometry, 2012, vol. 8, No. 1

Ähnliche Einträge