How Discrete Spectrum and Resonances Influence the Asymptotics of the Toda Shock Wave

We rigorously derive the long-time asymptotics of the Toda shock wave in a middle region where the solution is asymptotically a finite gap. In particular, we describe the influence of the discrete spectrum in the spectral gap on the shift of the phase in the theta-function representation for this so...

Full description

Saved in:
Bibliographic Details
Published in:Symmetry, Integrability and Geometry: Methods and Applications
Date:2021
Main Authors: Egorova, Iryna, Michor, Johanna
Format: Article
Language:English
Published: Інститут математики НАН України 2021
Online Access:https://nasplib.isofts.kiev.ua/handle/123456789/211304
Tags: Add Tag
No Tags, Be the first to tag this record!
Journal Title:Digital Library of Periodicals of National Academy of Sciences of Ukraine
Cite this:How Discrete Spectrum and Resonances Influence the Asymptotics of the Toda Shock Wave. Iryna Egorova and Johanna Michor. SIGMA 17 (2021), 045, 32 pages

Institution

Digital Library of Periodicals of National Academy of Sciences of Ukraine
_version_ 1860056769521254400
author Egorova, Iryna
Michor, Johanna
author_facet Egorova, Iryna
Michor, Johanna
citation_txt How Discrete Spectrum and Resonances Influence the Asymptotics of the Toda Shock Wave. Iryna Egorova and Johanna Michor. SIGMA 17 (2021), 045, 32 pages
collection DSpace DC
container_title Symmetry, Integrability and Geometry: Methods and Applications
description We rigorously derive the long-time asymptotics of the Toda shock wave in a middle region where the solution is asymptotically a finite gap. In particular, we describe the influence of the discrete spectrum in the spectral gap on the shift of the phase in the theta-function representation for this solution. We also study the effect of possible resonances at the endpoints of the gap on this phase. This paper is a continuation of research started in [arXiv:2001.05184].
first_indexed 2026-03-19T02:46:08Z
format Article
fulltext Symmetry, Integrability and Geometry: Methods and Applications SIGMA 17 (2021), 045, 32 pages How Discrete Spectrum and Resonances Influence the Asymptotics of the Toda Shock Wave Iryna EGOROVA a and Johanna MICHOR b a) B. Verkin Institute for Low Temperature Physics and Engineering, 47, Nauky Ave., 61103 Kharkiv, Ukraine E-mail: iraegorova@gmail.com b) Faculty of Mathematics, University of Vienna, Oskar-Morgenstern-Platz 1, 1090 Wien, Austria E-mail: Johanna.Michor@univie.ac.at URL: http://www.mat.univie.ac.at/~jmichor/ Received January 21, 2021, in final form April 26, 2021; Published online May 01, 2021 https://doi.org/10.3842/SIGMA.2021.045 Abstract. We rigorously derive the long-time asymptotics of the Toda shock wave in a middle region where the solution is asymptotically finite gap. In particular, we describe the influence of the discrete spectrum in the spectral gap on the shift of the phase in the theta-function representation for this solution. We also study the effect of possible resonances at the endpoints of the gap on this phase. This paper is a continuation of research started in [arXiv:2001.05184]. Key words: Toda equation; Riemann–Hilbert problem; steplike; shock 2020 Mathematics Subject Classification: 37K40; 35Q53; 37K45; 35Q15 1 Introduction The Toda shock wave describes the motion of an infinite chain of particles with nonlinear nearest neighbor interactions when the chain is excited with shock type initial conditions. We are interested in the effect the eigenvalues in the spectral gap of the associated Lax operator have on the asymptotic behavior of the shock wave. The Toda shock wave is generated by the solution of the following initial value problem for the Toda lattice [23, 24] d dt b̃(n, t) = 2 ( ã(n, t)2 − ã(n− 1, t)2 ) , d dt ã(n, t) = ã(n, t) ( b̃(n+ 1, t)− b̃(n, t) ) , (n, t) ∈ Z× R+, (1.1) with a steplike initial profile { ã(n, 0), b̃(n, 0) } such that ã(n, 0)→ a±, b̃(n, 0)→ b±, as n→ ±∞, (1.2) where a± > 0 and b± ∈ R satisfy the condition b− + 2a− < b+ − 2a+. (1.3) This condition fixes the position of the background spectra relative to each other; their mutual location produces essentially different types of asymptotic solutions [20]. The notion of the Toda shock wave [4, 5] was traditionally associated with symmetric initial data ã(n− 1, 0) = ã(−n, 0), b̃(n, 0) = −b̃(−n, 0), (1.4) mailto:iraegorova@gmail.com mailto:Johanna.Michor@univie.ac.at http://www.mat.univie.ac.at/~jmichor/ https://doi.org/10.3842/SIGMA.2021.045 2 I. Egorova and J. Michor and the background constants a− = a+ = 1 2 , b+ = −b− > 1. The asymptotic of the solution of (1.1) for the particular case ã(n, 0) = 1 2 , b̃(n, 0) = b sgnn, n ∈ Z, where sgn 0 = 0, (1.5) was studied in the pioneering work [25] by Venakides, Deift, and Oba in 1991. By use of the Lax– Levermore approach they established that in a middle region of the half plane (n, t) ∈ Z× R+, the asymptotic of the shock wave (1.1), (1.5) is described by a 2-periodic solution of the Toda lattice. They also showed that the asymptotic undergoes a phase shift caused by the presence of a single eigenvalue at λ = 0. We refer to this middle region of periodic asymptotics as VDO region,1 compare Figure 1. In this paper, we offer a derivation and rigorous justification of the asymptotic for (1.1)–(1.3) in the VDO region using the vector Riemann–Hilbert problem (RHP) approach. We allow more general initial data (1.2) with arbitrary positive a± and b± satisfying (1.3). In particular, the novel features are: � an arbitrary discrete spectrum, � possible resonances at the edges of the continuous spectrum, � no symmetry assumption (1.4), � a partial revision of results in [25] including estimates on the error terms, � a finite gap (two band) asymptotic due to spectra of different length. The vector RHP approach in the context of the Toda problem was proposed in [6] and further developed in [2, 12, 17, 18, 19]. We use standard conjugations/deformations such as the g-function technique [8] which proved its efficiency in steplike cases. A suitable g-function for the VDO region replaces the standard phase function and makes it possible to apply the lense mechanism. It also provides a characterization of the boundaries of the sectors (see Figure 2) where the asymptotics are given by a finite gap solution of (1.1) with unaltered phase. We des- cribe the g-function for the VDO region as an Abel integral on the Riemann surface associated with the continuous two band spectrum of the underlying Jacobi operator of (1.1) in Section 3. Before we state our main theorem, let us first note that without loss of generality it is sufficient to study the case of background spectra [b− 2a, b+ 2a] ∪ [−1, 1]. Indeed, assume that the vector-function ( ã(t), b̃(t) ) = { ã(n, t), b̃(n, t) } n∈Z is the solution of (1.1)–(1.3). Then the function (a(t), b(t)) given by a(n, t) = 1 2a+ ã ( n, t 2a+ ) , b(n, t) = 1 2a+ b̃ ( n, t 2a+ ) − b+, n ∈ Z, satisfies the initial value problem d dt b(n, t) = 2 ( a(n, t)2 − a(n− 1, t)2 ) , d dt a(n, t) = a(n, t)(b(n+ 1, t)− b(n, t)), (n, t) ∈ Z× R+, (1.6) a(n, 0)→ 1 2 , b(n, 0)→ 0, n→ +∞; a(n, 0)→ a, b(n, 0)→ b, n→ −∞, with b+ 2a < −1, (1.7) where we denoted b := b− − b+, a := a− 2a+ . 1Precise boundaries for the VDO region in our general case are given by (3.5)–(3.7). How Discrete Spectrum and Resonances Influence the Asymptotics 3 Hence it suffices to study the shock wave (1.6)–(1.7). We assume that the initial data tend to the background constants exponentially fast with some small rate ρ > 0, ∞∑ n=1 eρn (∣∣∣∣a(n, 0)− 1 2 ∣∣∣∣+ |b(n, 0)|+ |a(−n, 0)− a|+ |b(−n, 0)− b| ) <∞. (1.8) Figure 1 demonstrates the behavior of the Toda shock wave corresponding to the initial data a(n, 0) = 1 2 , n ∈ Z; b(n, 0) = −4, n < 0; b(0, 0) = −1.7, b(n, 0) = 0, n > 0, at a large but fixed time t = 200. Such initial data have one eigenvalue in the gap and the background spectra are of equal length. Hence the asymptotic of the shock wave in the VDO region is periodic with period 2 and exhibits one phase shift. In the left and right modulation regions (MR) the asymptotic is a modulated single-phase quasi-periodic Toda solution as discussed in [11]. ����� �� -��� -��� � ��� ��� ��� ��� ��� ��� �(�� ���) 0.5 1.0 1.5 2.0 a(n, 200) −400 −200 0 200 400 ��� ���� -��� -��� � ��� ��� -��� -��� -��� -��� -��� -��� -��� -��� ��� �(�� ���) −4.0 −3.5 −3.0 −2.5 −2.0 −1.5 −1.0 −0.5 0.0 b(n, 200) −400 −200 0 200 400 Figure 1. Numerically computed Toda shock wave with one eigenvalue. The initial data (1.8) can have a finite discrete spectrum. We enumerate the eigenvalues λj in the gap (b+2a,−1) increasingly starting from the leftmost; let ℵ be the number of eigenvalues in the gap. Given an arbitrary2 small ε > 0, the VDO region consists of ℵ+ 1 disjoint regions{ (n, t) : n t ∈ Ijε = [ξj + ε, ξj−1 − ε] } as depicted in Figure 2, where ξj are the points ξ at which the level line Re g(λ, ξ) = 0 of the g-function (cf. Section 3) crosses R at λj . We denote by ξ0 and ξℵ+1 the points where Re g(b + 2a, ξ0) = 0 and Re g(−1, ξℵ+1) = 0, and the rays n t = ξℵ+1 and n t = ξ0 determine 2The maximal value of ε which is admissible for our purpose is specified in Section 4. 4 I. Egorova and J. Michor n t n t ∈ I j ε n t ∈ I 1 ε n t = ξ0 n t = ξ0 − ε n t = ξ1 + ε n t = ξj−1 n t = ξj n t = ξℵ+1 Figure 2. The VDO region. the outer boundaries of the VDO region. With each interval Ijε we associate a shift phase ∆j (not depending on ε) expressed in terms of the initial scattering data for the solution of (1.6) and (1.8) (see (4.16) below). For each ∆j one finds via Jacobi’s inversion problem (3.12) the initial Dirichlet divisor and the unique finite gap solution { â(n, t,∆j), b̂(n, t,∆j) } from the isospectral set associated with the two-band spectrum S := [b− 2a, b+ 2a] ∪ [−1, 1]. Our main result is Theorem 1.1. Let {a(n, t), b(n, t)} be the solution of the initial value problem (1.6)–(1.7), (1.8) and let n → ∞, t → ∞ with n t ∈ I j ε , where ε > 0 is an arbitrary, sufficiently small number. Let { â(n, t,∆j), b̂(n, t,∆j) } be the finite gap solution associated with the spectrum S and with the phase ∆j given by (3.16), (4.16), (4.6), (4.12), (4.11), (2.23). Then there exists C(ε) > 0 such that b(n, t) = b̂(n, t,∆j) +O ( e−C(ε)t ) , a(n, t) = â(n, t,∆j) +O ( e−C(ε)t ) . (1.9) The shift of the phases at the point λj = zj+z −1 j 2 ∈ (b+2a,−1) of the discrete spectrum is given by ∆j −∆j+1 = 2 ∫ q q1 log ∣∣∣ zjs−1 s−zj ∣∣∣ ∣∣P̃−1(s) ∣∣ds∫ q1 −1 P̃−1(s) ds , P̃(s) = √ (s− q)(s− q1) ( s− q−1 1 )( s− q−1 ) , where q and q1 are defined by (2.6). Remark 1.2. (i) For initial data (1.8) the scattering data consist of the modulo of the right transmission coefficient |T (λ)| given on [b − 2a, b + 2a], the right reflection coefficient R(λ) on [−1, 1] and the discrete spectrum on R \ ([b− 2a, b+ 2a] ∪ [−1, 1]). As expected, we see that R(λ) and the discrete spectrum to the right of λj do not influence the asymptotic in the sector n t ∈ I j ε . (ii) The error terms in (1.9) are of order O ( e−C(ε)t ) and thus significantly better than the estimate O ( t−1 ) one would expect by analogy with the error estimates in the modulation re- gions [11]. The error terms in (1.9) were obtained by a careful analysis of the relations between the analytic continuation of the scattering functions. These relations allowed us to prove that there are no parametrix points [7] in the RHP for the VDO region. (iii) We use vector RHP statements instead of matrix statements (as do [1, 10, 14, 15, 22] in the case of the KdV equation with steplike initial data), because the matrix statements for the shock wave are ill-posed for certain arbitrary large values of n and t in the class of invertible How Discrete Spectrum and Resonances Influence the Asymptotics 5 matrices with L2-integrable singularities on the jump contour, for both the initial and model RHPs. This fact for Toda shock can be established similarly as for the KdV case [13]. One would have to admit then additional poles for solutions outside the discrete spectrum in the matrix statements. This makes proving uniqueness of the solutions far more difficult. The statements of the RHPs in vector form together with additional symmetries to be posed on contours, jump matrices and on the solutions itself imply uniqueness almost straightforward. However, for the final small-norm arguments we need to construct an invertible matrix model RHP solution. It has poles and it might not be unique, but the corresponding error vector function has no poles. Such a solution is given in Lemma 5.4. (iv) Unlike to KdV, for the Toda equation with non-overlapping background spectra, the statements of the RHPs associated with left and right initial data look identical. The proper choice of the initial statement for the RHP can essentially simplify the further analysis in a given region of space-time variables Z×R+ (cf. [12]). For the VDO region both choices are appropriate. 2 Notations and statement of the initial holomorphic RHP To maintain generality of the presentation while keeping notations short, we formulate all pre- liminary facts on the inverse scattering transform for the steplike initial profile in terms of the spectral variables z± associated with the initial data (1.2), (1.3). First of all, let us list some well known properties of the scattering data for the steplike Jacobi operator H̃(t) involved in the Lax representation d dtH̃(t) = [ H̃(t), Ã(t) ] for the initial value problem (1.1), (1.2). This problem has a unique solution (cf. [23]). Assume that the coefficients of the initial Jacobi operator H̃(0) tend to the limiting (or background) constants a±, b± with a first summable moment of pertur- bation, that is, n(ã(n, 0)−a±) ∈ `1(Z±) and n(b̃(n, 0)− b±) ∈ `1(Z±). Then the unique solution{ ã(t), b̃(t) } of (1.1) satisfies n ( ã(n, t)− a± ) ∈ `1(Z±), n ( b̃(n, t)− b± ) ∈ `1(Z±). (2.1) With this condition fulfilled, introduce some notations and notions. � The background Jacobi operators H±y(n) := a±y(n− 1) + b±y(n) + a±y(n+ 1), n ∈ Z, have spectra σ± = [b± − 2a±, b± + 2a±] which do not overlap, and by (1.3) satisfy supσ− < inf σ+. � The Joukovski maps z± = z±(λ) of the spectral parameter λ are given by λ = b± + a± ( z± + z−1 ± ) , z± : clos(C \ σ±) 7→ |z±| ≤ 1. The map z+ 7→ z− is one-to-one between the domains D+ and D−, where D± = {z± : |z±| < 1, z± /∈ z±(σ∓)}. The functions { z±n± } n∈Z are called the free exponents. They solve the background spectral equations H±y(n) = λy(n). � The operator H̃(t) has an absolutely continuous spectrum on the set σ+ ∪ σ− and a finite discrete spectrum σd, which we divide into three parts, σleft d = {λj ∈ σd : λj < b− − 2a−}, σright d = {λj ∈ σd : λj > b+ + 2a+}, σgap d = {λj ∈ σd : b− + 2a− < λj < b+ − 2a+}. The points z±j = z±(λj) ∈ D± ∩ (−1, 1), λj ∈ σd, are also called points of the discrete spectrum. 6 I. Egorova and J. Michor � The Jost solutions of the spectral equation ã(n− 1, t)ψ±(λ, n− 1, t) + ( b̃(n, t)− λ ) ψ±(λ, n, t) + ã(n, t)ψ±(λ, n+ 1, t) = 0 are normalised as lim n→±∞ (z±)∓nψ±(λ, n, t) = 1, λ ∈ clos(C \ σ±). We can consider them as functions of z± in the closures of D±. Their Wronskian W̃ (λ) := a(n− 1, 0) ( ψ−(λ, n− 1, 0)ψ+(λ, n, 0)− ψ+(λ, n− 1, 0)ψ−(λ, n, 0) ) (2.2) is an important spectral characteristic of the steplike scattering problem. It can be treated as an analytic function of z± in closD±. The points z±j are its simple zeros, and ψ±(λj , n, t) are the (dependent) eigenfunctions. � The normalising constants are introduced by(∑ n∈Z ψ2 ±(λj , n, t) )−2 = γ±j (t) = γ±j (0)ez ± j −(z±j )−1 . � The scattering relations T±(λ, t)ψ∓(λ, n, t) = R±(λ, t)ψ±(λ, n, t) + ψ±(λ, n, t) (2.3) hold on the sets |z±| = 1. � The time evolution of the scattering data is given by R±(λ, t) = R±(λ, 0)e±z±∓z −1 ± for |z±| = 1, |T±(λ, t)|2 = |T±(λ, 0)|2e±z∓∓z −1 ∓ for |z∓| = 1. (2.4) Relations (2.3) and (2.4) hold under condition (2.1) which guarantees existence and good ana- lytical properties of the Jost solutions. The function |T±(λ, 0)|2 cannot be continued analytically outside the domain |z∓| = 1. However, if the initial data tend to the limiting constants expo- nentially fast with a rate ρ > 0 (cf. (1.8)), then the right hand side of the ±-scattering relation continues analytically in the domain 1 − ρ < |z±| ≤ 1, and the respective equality (2.3) is pre- served. In particular, the reflection coefficient R±(λ, 0) continues in the domain 1−ρ < |z±| ≤ 1, and the function χ(λ) defined in (2.23) below can be continued analytically in both domains. The vector RHP connected with the scattering problem for H̃(t) can be stated in two ways, based either on the right or left scattering data. The correct choice of the scattering data which significantly simplifies the further analysis depends on the region of the (n, t) half-plane for which the asymptotic of (1.1)–(1.3) should be derived. In our situation, the VDO region on the Z×R+ half plane could be analysed via left or right RHP and both cases are equivalent in structure and complexity of steps. To state a proper vector RHP we proceed as follows. Let M be the two-sheeted Riemann surface associated with the function w(λ) = (( (λ− b−)2 − 4(a−)2 )( (λ− b+)2 − 4(a+)2 ))1/2 , with glued cuts along σ+ and σ−. Denote a point on M by p = (λ,±). On the upper sheet of M introduce two 1 × 2 vector-functions M±(p, n, t) := M±(p) (here variables n and t are treated as parameters) by M±(p) = ( T±(λ, t)ψ∓(λ, n, t)(z±(λ))n, ψ±(λ, n, t))(z±(λ))−n ) , p = (λ,+). How Discrete Spectrum and Resonances Influence the Asymptotics 7 The first component of each function is a meromorphic function of p on the upper sheet on M with simple poles at points of σd and known residues. At infinity M±(p) have finite values and the product of components is equal to 1 [12]. Let us extend each function M± to the lower sheet by M±(p∗) = M±(p)σ1, (2.5) where σ1 = ( 0 1 1 0 ) is the first Pauli matrix and p∗ = (λ,−) is the involution point for p = (λ,+). With this extension, both functions have jumps along the boundaries of the sheets on M, and these jumps can be easily evaluated. The jump problems together with normalisation conditions M±1 (∞+)M±1 (∞−) = 1, residue conditions at points of σd and σ∗d, and symmetry condition (2.5), form the content of the left and right RHPs associated with (1.1)–(1.3). In this paper, we use the traditional RHP statement based on the right scattering data in terms of the variable z+. As discussed in the introduction, we restrict ourselves to the case a+ = 1 2 , b+ = 0, a− = a > 0, b− = b, b+ 2a < −1, and denote by S = [b − 2a, b + 2a] ∪ [−1, 1] the continuous spectrum of the Jacobi operator involved in (1.6). To ease notations, we omit from here on the subscript “+” in the notations and set z(λ) := z+(λ), λ = z + z−1 2 , |z| ≤ 1, q = z(b− 2a), q1 = z(b+ 2a), zj := z+(λj), γj := γ+ j (0). (2.6) Remark 2.1. We use the formal notation zj ∈ σd, σgap d if λj = zj+z −1 j 2 ∈ σd, σgap d , respectively. Let us enumerate the points zj starting from σgap d , that is, −1 < zℵ < · · · < z1 < q1, ℵ = Cardσgap d . All remaining points of the discrete spectrum will lie outside of σgap d . Further notations are R(z) := R+(λ, 0) for |z| = 1, ψ(z, n, t) = ψ+(λ, n, t) for |z| ≤ 1, T (z, t) := T+(λ, t), ψleft(z, n, t) = ψ−(λ(z), n, t) for z ∈ D, (2.7) where D = {z : |z| < 1, z /∈ [q1, q]}. (2.8) The domain D is in one-to-one correspondence with the upper sheet of M (we treat sheets as open sets) with z = λ− √ λ2 − 1↔ p = (λ,+), z ∈ D. The domain D∗ = { z : z−1 ∈ D } (2.9) corresponds to the lower sheet by z−1 = λ+ √ λ2 − 1↔ p∗ = (λ,−). 8 I. Egorova and J. Michor Therefore, the meromorphic RHP for M+(p) on M can be reformulated as an equivalent mero- morphic RHP for m(z) = M+(p(z)) on the z-plane, with jumps along the unit circle T = {z : |z| = 1} and intervals [q1, q] and [ q−1, q−1 1 ] . In this paper, we propose a slightly different (holomorphic) statement of the initial RHP, which is equivalent to the RHP for M+(p) on M, and therefore has a unique solution (cf. [12]). This statement is specific for the domain VDO, where we derive the asymptotics, and allows us to skip several of the standard transformations, such as the reformulation of the meromorphic problem as a holomorphic problem and one of two steps corresponding to opening of lenses. Let us choose a large natural number N � 1 and set δ ≤ N−1 min zi 6=zj∈σd { |zi − q|, |zi − zj |, |zi − q1|, ||zi| − 1| } . (2.10) We can always assume that δ < ρ, where ρ is the decay rate from (1.8). Then the right Jost solutions ψ(z, n, t) and ψ(z−1, n, t) are holomorphic functions in an δ-vicinity of the unit circle T, and the standard scattering relation T (z, t)ψleft(z, n, t) = ψ ( z−1, n, t ) +R(z, t)ψ(z, n, t) is continued analytically in the open ring Ω̃δ = {z : 1− δ < |z| < 1}. With our choice of δ, there are no points of the discrete spectrum in Ω̃δ, moreover, inf zj∈σd dist(zj , Cδ) > (N − 1)δ, where we denoted Cδ = {z : |z| = 1− δ}. (2.11) In particular, the continuation of the initial reflection coefficient R(z) is an analytic function in Ω̃δ. Set Dδ,j = {z : |z − zj | < δ}, Tδ,j = {z : |z − zj | = δ}, D∗δ,j = { z : z−1 ∈ Dδ,j } , Dδ = D \ ( Ω̃δ ∪ ⋃ σd Dδ,j ) , D∗δ = { z : z−1 ∈ Dδ } . (2.12) In D \ Σδ, where Σδ = I ∪ Cδ ∪ ⋃ σd Tδ,j , (2.13) with I := [q1, q], introduce the vector-function m(z) = (m1(z, n, t),m2(z, n, t)) by3 m(z) =  ( T (z, t)ψleft(z, n, t)z n, ψ(z, n, t)z−n ) , z ∈ Dδ,( ψ(z−1, n, t)zn, ψ(z, n, t)z−n ) , z ∈ Ω̃δ,( T (z, t)ψleft(z, n, t)z n, ψ(z, n, t)z−n ) Aj(z), z ∈ Dδ,j . (2.14) Here Aj(z) = ( 1 0 (z − zj)−1γjz 2n+1 j et(zj−z −1 j ) 1 ) . (2.15) 3This is a function of z, and n and t are treated as large parameters. How Discrete Spectrum and Resonances Influence the Asymptotics 9 Lemma 2.2 ([12]). We have m1(0, n, t) = ∞∏ k=n 2a(k, t), lim z→0 1 2z (m1(z, n, t)m2(z, n, t)− 1) = b(n, t). (2.16) Extend m(z) to D∗ \ Σ∗δ with Σ∗δ = { z : z−1 ∈ Σδ } by m(z−1) = m(z)σ1. (2.17) Formula (2.17) implies that the vector function (2.14), considered as a piecewise-analytic function in C, has jumps along the circle Cδ, along the interval I and the small circles Tδ,j , as well as along their images C∗δ , I∗ and T∗δ,j under the map z → z−1. However, m(z) does not have a jump along the unit circle |z| = 1, i.e., it is holomorphic in the ring 1− δ < |z| < (1− δ)−1. The fact that m(z) does not have any singularities at zj ∈ σd is established in [6] and [19]. The symmetry condition (2.17) plays a crucial role in establishing uniqueness of the solution for RHPs, and we cannot violate it. For this reason, the initial RHP and all its further transfor- mations (deformations and conjugations) should satisfy the following symmetry constraints (i) and (ii). Let Σ be the jump contour of a generic RHP. (i) The jump contour Σ should be symmetric with respect to the map z 7→ z−1, i.e., with every point z it also contains z−1. (ii) Symmetric parts of Σ are oriented in such a way that the jump matrix ṽ(z) of the problem m̃+(z) = m̃−(z)ṽ(z) and the solution itself satisfy the symmetries ṽ(z) = σ1ṽ ( z−1 ) σ1, z ∈ Σ; m̃(z) = m̃ ( z−1 ) σ1, z ∈ C \ Σ. Constraint (ii) implies that the orientation of symmetric parts of Σ is as follows: if a point z moves along a part of the contour K ⊂ Σ ∩ {z : |z| < 1} in the positive direction, then the point z−1 moves simultaneously in the positive direction of the symmetric part K∗, where K ∩ K∗ = ∅. Except for the lense mechanisms where triangle matrices are used, all conjugations of the solution vector consist of multiplication by diagonal matrices of the form [d(z)]−σ3 , that is, in transformations m̃(z) 7→ m̃(z)[d(z)]−σ3 , where d : C \Σ→ C is a sectionally analytic function and σ3 = ( 1 0 0 −1 ). On all such conjugations we pose the symmetry constraint (iii): (iii) The contour Σ of a non-analyticity for d(z) should be symmetric with respect to z 7→ z−1. Moreover, the function d(z) should satisfy either the property d ( z−1 ) = d(z)−1, z ∈ C \ Σ, or the property d ( z−1 ) = d(z) 6= 0, z ∈ C \ Σ, d(0) = 1. Recall that m(z) in (2.14) has bounded positive limits of both components at 0 and∞, moreover, by Lemma 2.2, m1(0)m2(0) = 1, m1(0) > 0. (2.18) This is a normalization condition. The properties of the conjugation matrices [d(z)]−σ3 listed above allow to preserve the normalisation condition for all transformations. The phase function of our problem is given by Φ(z) = Φ ( z, n t ) = 1 2 ( z − z−1 ) + n t log z, z ∈ clos(C \ R−). 10 I. Egorova and J. Michor Set ξ = n t and consider the cross points of the 0-level lines for the function Re Φ(z, ξ), that is, the lines described by 1 2 ( z − z−1 ) + ξ log |z| = 0, for different values of ξ ∈ R. One of the level lines for all ξ ∈ R is evidently the unit circle |z| = 1. If ξ ≥ 1 the other two lines are located in the domains |z| < 1 and |z| > 1 and are symmetric with respect to the map z 7→ z−1. An elementary analysis shows that the point −1 < z0(ξ) < 0, where the respective level line crosses the real axis moves monotonously from −1 to 0 when ξ runs the interval [1,+∞). Thus, the points z0(ξ) and z−1 0 (ξ) meet at point −1 for ξ = 1. The same analysis shows that for ξ ∈ [−∞, 0) the cross point z0(ξ) moves monotonously along the interval [0, 1] and z0(1) = z−1 0 (1) = 1. When ξ ∈ [−1, 1], the points z0(ξ) and z−1 0 (ξ) = z0(ξ) lie on the unit circle. To state the RHP for which m(z) in (2.14) is the unique solution,4 we introduce orientations on the jump contour Σδ ∪Σ∗δ according to the symmetry requirements above. The contour Cδ is oriented counterclockwise, C∗δ is oriented clockwise. On the two symmetric parts I := [q, q1], I∗ := [ q−1, q−1 1 ] , (2.19) the orientation5 is taken from right to left on I and from left to right on I∗. Moreover, all Tδ,j and T∗δ,j are supposed to be oriented counterclockwise. Then m+(z) (resp. m−(z)) will denote the limit from the positive (resp. negative) side of the contour. We assume that these limits exist and m(z) extends to a continuous function on the sides of Σδ ∪ Σ∗δ except possibly at the end points of I and I∗, J := { q, q1, q −1, q−1 1 } , (2.20) where the square root (not L2-integrable!) singularities are admissible. Recall that m1(z) = m2 ( z−1 ) = O(z − q̃)−1/2 as z → q̃ ∈ {q, q1} iff W (q̃) = 0, where W (z) = W̃ (λ(z)) (cf. (2.2)) is the Wronskian of the Jost solutions, z ∈ closD. If the Wronskian vanishes at q̃, we call q̃ a resonant point. The general situation is non-resonant, that is, W (q̃) 6= 0. Note that as a function of z, the Wronskian takes complex conjugated values on the sides of the contours (2.19). T∗δ,j σ1v ( z−1 ) σ1 Tδ,j Aj(z) Cδ C∗δ I∗ σ1v ( z−1 ) σ1 I ( 1 0 χ(z)e2tΦ(z) 1 ) ( 1 0 R(z)e2tΦ(z) 1 ) σ1v ( z−1 ) σ1 Figure 3. Jump matrix v(z) in Theorem 2.3. 4We formulate a RHP which is equivalent to the initial RHPs considered in [11] or [12] in the domain under consideration. Uniqueness is proven in [12]. 5In what follows, for a < b the notation [b, a] means that the contour is oriented from b to a. How Discrete Spectrum and Resonances Influence the Asymptotics 11 Theorem 2.3 ([11, 12]). Let δ > 0 be given as in (2.10). For all (n, t) ∈ Z × R+ and z ∈ C, the function (2.14), (2.17) is the unique solution of the following RH problem: to find a vector- function m(z) = (m1(z),m2(z)) holomorphic in C \ (Σδ ∪ Σ∗δ) which � is continuous up to the boundary, except of possibly points in J (2.20), � satisfies conditions (2.17), (2.18) and the jump condition m+(z) = m−(z)v(z), z ∈ Σδ ∪ Σ∗δ , (2.21) v(z) =  ( 1 0 χ(z)e2tΦ(z) 1 ) , z ∈ I,( 1 0 R(z)e2tΦ(z) 1 ) , z ∈ Cδ, Aj(z), z ∈ Tδ,j , zj ∈ σd, σ1v(z−1)σ1, z ∈ Σ∗δ , (2.22) where Φ(z) = Φ(z, nt ) is the phase function, matrices Aj(z) are defined by (2.15) and R(z) is the holomorphic continuation of the initial right reflection coefficient (2.7). The func- tion χ(z) is defined by χ(z) := ( z − z−1 )[ ζ−1 − ζ ] (z − i0) |W (z)|2 , z ∈ I, (2.23) with ζ(z), z ∈ D, connected with z by the Joukovski map λ = b + a(ζ + ζ−1). Here W (z) = W̃ (λ(z)) is the Wronskian of the Jost solutions defined in (2.2), � in vicinities of the points in J , m(z) has the following behavior: – if q̃ ∈ {q, q1} is nonresonant, then m(z) = (O(1), O(1)) as z → q̃, z → q̃−1, – if q̃ ∈ {q, q1} is resonant, i.e., if χ(z) = O ( 1√ z−q̃ ) as z → q̃, then m(z) = ( O ( 1√ z − q̃ ) , O(1) ) , z → q̃, m(z) = ( O(1), O ( 1√ z − q̃−1 )) , z → q̃−1. Remark 2.4. According to (1.8), W (z) admits an analytic continuation in a small vicinity of the interval I (see [11, equation (2.11)]). Respectively, in this vicinity there exists an analytic continuation X(z) of χ(z) such that X±(z) = ±i|χ(z)| for z ∈ I. The function X(z) does not have other jumps in this vicinity. Note that in the VDO region, the off-diagonal matrix elements of the jump matrix v(z) grow exponentially with respect to t for z ∈ I∪I∗. The same is true for v(z) on those contours Tδ,j∪T∗δ,j which correspond to the origins 0 > zj > z0(ξ). The remaining parts of the jump matrices are asymptotically close to the identity matrix as t→∞. In the next sections we perform a series of conjugation/deformation steps which transform the initial RHP of Theorem 2.3 to the equivalent problem with a jump matrix which is asymptotically close as t → ∞ to a piecewise constant matrix with respect to z. This limiting matrix also depends on ξ as a piecewise constant matrix, and the respective (so called model) RHP has a unique solution which can be found explicitly in terms of the Riemann theta-function. Let us emphasize that for the VDO region we propose transformations which lead to the absence of any additional parametrix problems. The first 12 I. Egorova and J. Michor transformation of the initial problem solution is associated with the so called g-function method first introduced for the KdV equation in [8]. In our case, this g-function is a normalized Abel integral associated with the two-sheeted Riemann surface glued via the cuts along the continuous spectrum S = [b−2a, b+2a]∪ [−1, 1]. In fact, the g-function is a linear combination (dependent on ξ) of the normalized Abel differentials of the second and third type which are involved in the exponential part of the Baker–Akhiezer function corresponding to the finite gap solution of the Toda lattice associated with the two-band spectrum S. It is crucial for our endeavor to understand in detail the properties of the g-function as a function of the spectral parameter λ. The next section is devoted to this subject. 3 g-function as an Abel integral and its connection with the Baker–Akhiezer function Let M be the two-sheeted Riemann surface associated with S, i.e., with the function R1/2(λ) = − √( λ2 − 1 )( (λ− b)2 − 4a2 ) , with sheets ΠU and ΠL glued along the cuts over the intervals [b − 2a, b + 2a] and [−1, 1]. The indices U and L label the upper and lower sheets of the surface. We denote by p = (λ,±) the points of M, with (∞,±) :=∞±; and p∗ = (λ,∓) for p = (λ,±) denotes the sheet exchange map. Choose a canonical basis of a and b cycles on M as follows: the cycle b surrounds the interval [b− 2a, b+ 2a] counterclockwise on ΠU and the cycle a passes from b+ 2a to −1 on the upper sheet and back on the lower sheet. The part of this cycle on ΠU we denote by JU and treat it as a contour on M. Its projection lies on the interval [b+ 2a,−1] which we call the gap. The lower part of a has the same projection on the gap, and is considered as the contour JL which passes from −1 to b+ 2a on ΠL. Let Ω0 be the Abel differential of the second kind on M with second order poles at ∞+ and∞− and let ω∞+,∞− be the Abel differential of the third kind with logarithmic poles at∞+ and ∞−, both normalized as∫ a Ω0 = ∫ a ω∞+,∞− = 0. As it is known, Ω0 = (λ− ν1)(λ− ν2) R1/2(λ) dλ, ω∞+,∞− = λ− ν3 R1/2(λ) dλ, (3.1) where νi ∈ R for i = 1, 2, 3. Moreover, ν3 ∈ (b+ 2a,−1), and at least one of the points ν1 or ν2 also lies in the gap (b+ 2a,−1). Consider the Abel integral given by g(p, ξ) := ∫ p 1 Ω0 + ξ ∫ p 1 ω∞+,∞− , where ξ ∈ R is a parameter. On ΠU we denote it by g(λ, ξ), that is, g(λ, ξ) = ∫ λ 1 (λ− ν1)(λ− ν2)− ξ(λ− ν3) R1/2(λ) dλ = ∫ λ 1 (λ− µ1(ξ))(λ− µ2(ξ)) R1/2(λ) dλ. (3.2) Since ∫ −1 b+2a (λ− µ1(ξ))(λ− µ2(ξ)) R1/2(λ) dλ = 0, (3.3) How Discrete Spectrum and Resonances Influence the Asymptotics 13 then µi(ξ) ∈ R, these points do not coincide, and at least one of them belongs to the gap. By definition of Ω0 we have (λ− ν1)(λ− ν2) R1/2(λ) = −1 +O ( λ−2 ) , that is, (λ− µ1(ξ))(λ− µ2(ξ)) R1/2(λ) + 1 = −µ1(ξ)− µ2(ξ) + b λ +O ( λ−2 ) = ξ λ +O ( λ−2 ) , which implies µ1(ξ) = b− ξ − µ2(ξ). (3.4) By (3.3),∫ −1 b+2a (λ− µi(ξ))(λ− b+ ξ + µi(ξ))) R1/2(λ) dλ = 0, that is, µi(ξ) are the zeros of the quadratic equation µ2 ∫ a dλ R1/2(λ) + µ(ξ − b) ∫ a dλ R1/2(λ) − ∫ a λ2 + λ(ξ − b) R1/2(λ) dλ = 0. With the notations Γ1 = ∫ a λ2dλ R1/2(λ)∫ a dλ R1/2(λ) , Γ2 = ∫ a λdλ R1/2(λ)∫ a dλ R1/2(λ) , (3.5) we infer µ1,2(ξ) = 1 2 ( b− ξ ± √ (b− ξ)2 + 4(Γ1 + (ξ − b)Γ2) ) . (3.6) Lemma 3.1. The functions µi(ξ), i = 1, 2, are monotonically decreasing with respect to ξ ∈ R. For ξ ∈ (ξℵ+1, ξ0),6 where ξℵ+1 = b+ 1− Γ1 1 + Γ2 , ξ0 = b+ Γ1 − (b+ 2a)2 b+ 2a− Γ2 , (3.7) we have µi(ξ) ∈ (b+ 2a,−1), i = 1, 2. Proof. Differentiating (3.6) with respect to ξ implies 2 d dξ µi(ξ) = −1± ξ − b+ 2Γ2√ (b− ξ)2 + 4(Γ1 + (ξ − b)Γ2) . We observe that Γ1 > 1, Γ2 < −1, |Γ1| > |Γ2|. Inequality |ξ − b+ 2Γ2| < √ (b− ξ)2 + 4(Γ1 + (ξ − b)Γ2) 6In [12] and [11], these values were denoted by ξ′cr,1 and ξ′cr. 14 I. Egorova and J. Michor holds if Γ2 2 < Γ1. The last one follows from the Cauchy inequality∣∣∣∣ ∫ a λ dλ R1/2(λ) ∣∣∣∣ < √∫ a λ2dλ R1/2(λ) √∫ a dλ R1/2(λ) . Thus, µi(ξ) are monotonically decreasing with respect to ξ. Assume that µ1(ξ) < µ2(ξ). A trivial analysis shows that the value ξ0 corresponds to the location µ1(ξ0) = b+ 2a, that is, 2(b+ 2a) = b− ξ0 − √ (b− ξ0)2 + 4(Γ1 + (ξ0 − b)Γ2). This implies the second equation in (3.7). The location µ2(ξℵ+1) = −1 provides the first formula in (3.7). From (3.4) it follows that ξ0 = −2a− µ1(ξ0), ξℵ+1 = b+ 1− µ2(ξℵ+1), that is, ξ0 − ξℵ+1 = |b+ 2a| − 1 + µ2(ξℵ+1)− µ1(ξ0) > 0, since µ1(ξ0), µ2(ξℵ+1) ∈ (b+ 2a,−1). � Let ε > 0 be an arbitrary small number and let ξ0 and ξℵ+1 be defined by (3.7) and (3.5). For any ξ ∈ [ξℵ+1 + ε, ξ0 − ε], both points µ1(ξ) and µ2(ξ) are inner points of the gap. The set of level lines Re g = 0 consists of the two intervals [b − 2a, b + 2a] and [−1, 1] and an infinite contour which intersects the real axis at µ0(ξ) such that µ1(ξ) < µ0(ξ) < µ2(ξ). (3.8) Lemma 3.2. The real-valued function µ0(ξ) implicitly given by Re g(µ0(ξ), ξ) = 0 is monotonic with d dξµ0(ξ) < 0 for ξ ∈ (ξℵ+1, ξ0). Moreover, lim ξ→ξ0 µ0(ξ) = lim ξ→ξ0 µ1(ξ) = b+ 2a, lim ξ→ξℵ+1 µ0(ξ) = lim ξ→ξℵ+1 µ2(ξ) = −1. (3.9) Proof. By (3.6) we have µ1(ξ)µ2(ξ) = (b − ξ)Γ2 − Γ1. This implies with (3.4) that µ0(ξ) is given implicitly by∫ −1 µ0(ξ) (λ− µ1(ξ))(λ− µ2(ξ)) R1/2(λ) dλ = ∫ −1 µ0(ξ) λ2 + (ξ − b)λ+ ((b− ξ)Γ2 − Γ1) R1/2(λ) dλ = 0. Differentiating with respect to ξ implies d dξ µ0 = ∫ −1 µ0 λ− Γ2 R1/2(λ) dλ R1/2(µ0) (µ0 − µ1)(µ0 − µ2) . Since the second multiplier is negative, it is sufficient to prove that the integral is positive. But∫ −1 µ0 λ− Γ2 R1/2(λ) dλ = ∫ −1 µ0 λ R1/2(λ) dλ− ∫ −1 µ0 dλ R1/2(λ) ∫ −1 b+2a λ R1/2(λ) dλ (∫ −1 b+2a dλ R1/2(λ) )−1 , that is, we have to prove that∫ −1 µ0 λ dλ R1/2(λ)∫ −1 µ0 dλ R1/2(λ) > ∫ −1 b+2a λ dλ R1/2(λ)∫ −1 b+2a dλ R1/2(λ) . This inequality is true by the mean value theorem. Equalities (3.9) were proven in [12]. � How Discrete Spectrum and Resonances Influence the Asymptotics 15 Let us recall the Baker–Akhiezer function for a finite gap solution { â(n, t), b̂(n, t) } of the Toda lattice equation associated with the spectrum S = [b− 2a, b+ 2a] ∪ [−1, 1] and the initial Dirichlet divisor p0 = (λ(0, 0), σ(0, 0)), σ(0, 0) ∈ {+,−}. In our case the divisor consists of one point on the Riemann surface M with projection on the closed gap of the spectrum and it will later depend on the slow variable ξ. Let ζ be the holomorphic Abel differential on M normalized as ∫ a ζ = 1 and let ∫ b ζ =: τ ∈ iR+ be its b-period. Introduce the Abel map A(p) := ∫ p b−2a ζ. It is an odd function on M, A(p∗) = −A(p). Moreover, it has a jump along the a-cycle, which we interpret as a union JU ∪ JL. Then A+(p)−A−(p) = −τ as p ∈ JU ∪ JL. We set A(b+ 2a) := A+(b+ 2a) = −τ/2 and A(p0) = A+(p0), where p0 ∈ JU ∪ JL is the initial Dirichlet divisor. Let Ξ = τ 2 + 1 2 be the Riemann constant and Λ = ∫ b ω∞+,∞− ∈ iR, U = ∫ b Ω0 ∈ iR, (3.10) be the b-periods of the Abel differentials (3.1). Following [23], we introduce the notations Z(p, n, t) := A(p)−A(p0)− n Λ 2πi − t U 2πi − Ξ, Z(n, t) := Z(∞+, n, t). Evidently, θ(Z(p, 0, 0)) = 0 iff p = p0, where θ(v) := θ(v | τ) = ∑ m∈Z exp ( πim2τ + 2πimv ) is the Jacobi theta function. Recall that the time-dependent Baker–Akhiezer function for the finite gap Toda lattice solution with spectral data as above has the form [23] Ψ(p, n, t) = C(n, t) θ(Z(p, n, t)) θ(Z(p, 0, 0)) exp ( n ∫ p 1 ω∞+,∞− + t ∫ p 1 Ω0 ) = C(n, t) θ(Z(p, n, t)) θ(Z(p, 0, 0)) exp(tg(p, ξ)), ξ = n t . Here C(n, t) is a positive constant (with respect to p) which provides the equalities lim p→∞± Ψ(p∗, n, t)Ψ(p, n, t) = 1, Ψ(p, 0, 0) = 1, and C(n+ 1, t) C(n, t) = √ θ(Z(n− 1, t)) θ(Z(n+ 1, t)) > 0. As is known, for each n and t fixed, the Baker–Akhiezer function is a meromorphic function of p on M with a simple pole at p0. Respectively, the vector function (Ψ(p∗, n, t),Ψ(p, n, t)) does not have jumps on M, and the vector function m̂(p) := ( Ψ(p∗, n, t) exp ( −tg ( p∗, n t )) ,Ψ(p, n, t) exp ( −tg ( p, n t ))) , has an evident jump along a, m̂+(p) = m̂−(p)e−(nΛ+tU)σ3 , p ∈ JU ∪ JL. 16 I. Egorova and J. Michor Here we took into account that for p ∈ JU ∪ JL,[ ∫ p 1 ω∞+,∞− ] + − [ ∫ p 1 ω∞+,∞− ] − = −Λ, [ ∫ p 1 Ω0 ] + − [ ∫ p 1 Ω0 ] − = −U. Note that since g(p∗, ξ) = −g(p, ξ), m̂1(∞+)m̂2(∞+) = 1, m̂(p) = m̂(p∗)σ1. The function f(p) := f∞ θ(Z(p, 0, 0)) θ ( A(p)− 1 2 ) , f∞ := √ θ ( A(∞+)− 1 2 ) θ ( A(∞−)− 1 2 ) θ(Z(∞+, 0, 0))θ(Z(∞−, 0, 0)) , has a simple pole at the branch point b+ 2a and a simple zero at p0. Moreover, f(∞+)f(∞−) = 1, f+(p∗) = f−(p∗)e−i∆, f+(p) = f−(p)ei∆, p ∈ JU , where ∆ = −2π ( A(p0) + τ 2 ) ∈ R. (3.11) Note that (3.11) can be rewritten as the Jacobi inversion problem∫ p0 b+2a ζ = −∆ 2π (mod 1) (3.12) and allows us to compute uniquely the divisor point p0 for any given real valued ∆. Summing up the considerations above, we proved the following Lemma 3.3. The vector function m̃(p) = ( m̃1(p, n, t), m̃2(p, n, t) ) = ( Ψ(p∗, n, t)f(p∗) exp ( −tg ( p∗, n t )) ,Ψ(p, n, t)f(p) exp ( −tg ( p, n t ))) = C̃(n, t) ( θ(Z(p∗, n, t)) θ ( A(p∗)− 1 2 ) , θ(Z(p, n, t)) θ ( A(p)− 1 2 )), where C̃(n, t) := C(n, t)f∞, solves the following RHP on M: to find a holomorphic vector-function m̃(p) on M \ ( JU ∪ JL ) , which satisfies � the jump condition m̃+(p) = m̃−(p)e−(nΛ+tU+i∆)σ3 , (3.13) � the symmetry condition m̃(p∗) = m̃(p)σ1 for p ∈M \ ( JU ∪ JL ) , (3.14) � the normalization condition m̃1(∞+)m̃2(∞+) = 1. � Both components of m̃(p) have simple poles at the branch point p = b + 2a and no other singularities. How Discrete Spectrum and Resonances Influence the Asymptotics 17 Note that the constant C̃(n, t) in Lemma 3.3 satisfies C̃(n+ 1, t) C̃(n, t) = √ θ(Z(n− 1, t)) θ(Z(n+ 1, t)) > 0. Together with Theorem 9.48 of [23] it implies Corollary 3.4. For the vector function m̃(p) = m̃(p, n, t) the following holds m̃1(∞+, n, t) m̃1(∞+, n+ 1, t) = √ θ(Z(n− 1, t))θ(Z(n+ 1, t)) θ(Z(n, t)) = â(n, t) ã , where ã = CapS is the logarithmic capacity of the spectrum S. Introduce the product h̃(p) := m̃1(p)m̃2(p) = C̃2(n, t) θ(Z(p∗, n, t)) θ ( A(p∗)− 1 2 ) θ(Z(p, n, t)) θ ( A(p)− 1 2 ) , and let h(λ) := h̃(p) for p = (λ,+). The function h̃(p) has a double pole on M at the branch point b+ 2a, that is, h(λ) has a simple pole at b+ 2a. Moreover, θ(Z(p, n, t)) has the only zero at p(n, t) = (λ(n, t),±) ∈ JU ∪ JL, which is the unique solution of the Jacobi inversion problem∫ p(n,t) p0 ζ = n Λ 2πi + t U 2πi , (3.15) θ(Z(p∗, n, t)) has a simple zero at the involution point p∗(n, t). Thus h(λ(n, t)) = 0, and it is a simple zero of h. We observe that from the jump and symmetry conditions it follows that h̃(p) does not have jumps on M, moreover, h̃(p) = h̃(p∗), p ∈M. This means that h(λ) does not have jumps along the spectrum S and on the gap [b + 2a,−1]. The normalisation condition implies limλ→∞ h(λ) = 1. Hence h(λ) is a meromorphic function on C, i.e., h(λ) = λ−λ(n,t) λ−b−2a . Corollary 3.5. Let λ(n, t) ∈ [b + 2a,−1] be the projection on C of the Dirichlet eigenvalue p(n, t) given by (3.15). Then lim p→∞+ p (m̃1(p)m̃2(p)− 1) = b+ 2a− λ(n, t). We recall that the trace formula in our case looks like b̂(n, t) = 1 2 (1 + b− 2a+ b+ 2a− 1− 2λ(n, t)) = b− λ(n, t). Therefore, lim p→∞+ p (m̃1(p)m̃2(p)− 1) = b̂(n, t) + 2a. Since the problem (3.12) has a unique solution p0 for any real ∆, we can treat ∆ as the initial data to choose the representative { â(n, t), b̂(n, t) } for the isospectral set of finite gap potentials with spectrum S. To emphasise this dependence we denote the representative as { â(n, t,∆), b̂(n, t,∆) } . In turn, the solution of the RHP with jump (3.13) we denote as m̃(p, n, t,∆). We proved the following 18 I. Egorova and J. Michor Theorem 3.6. Let m̃(p, n, t,∆) be the unique solution of the RHP in Lemma 3.3. Then lim p→∞+ ã m̃1(p, n, t,∆) m̃1(p, n+ 1, t,∆) = â(n, t,∆), lim p→∞+ p (m̃1(p, n, t,∆)m̃2(p, n, t,∆)− 1)− 2a = b̂(n, t,∆), where { â(n, t,∆), b̂(n, t,∆) } is the finite gap solution of (1.6) with two band spectrum S = [b− 2a, b+ 2a] ∪ [−1, 1] and initial Dirichlet divisor p0 given by the Jacobi inversion (3.12). Remark 3.7. For convenience of the reader, we recall from [23] that the finite gap solution {â(n, t,∆), b̂(n, t,∆)} corresponding to the initial phase ∆ and spectrum S is given by â(n, t,∆) = ã √ θ ( (n−1)Λ 2πi + tU 2πi − ∆ 2π +A ) θ ( (n+1)Λ 2πi + tU 2πi − ∆ 2π +A ) θ ( nΛ 2πi + tU 2πi − ∆ 2π +A ) , b̂(n, t,∆) = b̃+ 1 Y ∂ ∂w log θ ( (n−1)Λ 2πi + tU 2πi − ∆ 2π +A+ w ) θ ( nΛ 2πi + tU 2πi − ∆ 2π +A+ w ) , (3.16) where A := 1 2 + ∫ b−2a ∞+ ζ, Y = ∫ a dλ R1/2(λ) , b̃ = b− 1 Y ∫ a λ dλ R1/2(λ) , ζ is the normalized holomorphic Abel differential, ã = CapS, and U , Λ are defined by (3.10). 4 Reduction of the initial RHP to the model RHP From now we work again in the variable z. Let us identify the upper sheet ΠU of the Riemann surface M with the domain (2.8) and the lower sheet with (2.9). The image of JU under the map p 7→ z we denote by J , preserving the orientation, i.e., J = [q1,−1] is oriented from right to left. As for the image J∗ of JL, we change its orientation in accordance with our symmetry requirements, i.e., J∗ = [ q−1 1 ,−1 ] is oriented from left to right. The other contours used here are already defined by (2.19), (2.11), (2.12), (2.13). In this section, we perform two transformations (steps) which transform the solution of the initial RHP (Theorem 2.3) to the solution of the RHP with the jump matrix which is close as t→∞ to a piecewise constant jump matrix everywhere on the jump contour, without exceptional points (parametrices). Step 1. Set yk = yk(ξ) = z(µk(ξ)) ∈ J for k = 0, 1, 2, where µk(ξ) are defined in Section 3. From (3.8) we have y2(ξ) < y0(ξ) < y1(ξ). With these definitions at hand, the g(p)-function (3.2) is given in terms of z by g(z, ξ) = 1 2 ∫ z 1 (s− y1) ( s− y−1 1 ) (s− y2) ( s− y−1 2 )√ (s− q1) ( s− q−1 1 ) (s− q) ( s− q−1 ) ds s2 , z ∈ C \ (−∞, 1). Thus, the level lines Re g(z, ξ) = 0 which are different from the unit circle T and intervals I and I∗, cross the real axis at the points y0(ξ) and y−1 0 (ξ). Similar to [11, Lemma 3.1] and [12, Lemma 5.3] we establish the following How Discrete Spectrum and Resonances Influence the Asymptotics 19 Lemma 4.1. The function g(z) satisfies the following properties (a) g(z) is single valued on C \ [ q−1, q ] and g ( z−1 ) = −g(z) for z ∈ C \ [ q−1, q ] , (b) Re g(z) = 0 for z ∈ I ∪ I∗ ∪ {z : |z| = 1}, (c) g(q) = g ( q−1 ) = 0, (d) g−(z) = −g+(z) for z ∈ I ∪ I∗, (e) Φ(z)− g(z) = K(ξ) +O(z) as z → 0, where K(ξ) ∈ R and d dξ K(ξ) = − log (2ã) . (4.1) Here ã is the logarithmic capacity of the set S, cf. [12, Lemma 5.4], (f) g+(z)− g−(z) = −U − ξΛ for z ∈ J , and g+(z)− g−(z) = U + ξΛ for J∗, where U and Λ are defined by (3.10). The signature table for Re g(z, ξ) is given in Figure 4. T∗ δ,j Tδ,j Re g < 0 Re g < 0Re g > 0 T qq1 0−1q−1 1q−1 y−1 0 y0 Figure 4. Signature table of Re g(z, ξ) for ξ ∈ Iε. Recall that we enumerated the eigenvalues of the problem (1.6), (1.8) starting from the gap (b + 2a,−1) in ascending order, i.e., λ1 is the minimal eigenvalue and λℵ is the maximal one. The respective zj are enumerated in descending order. Choose a small ε1 > 0 such that y1(ξ0 − ε1) > z1 + 2δ, y2(ξℵ+1 + ε1) < zℵ − 2δ. Smaller values of ε or δ do not affect these conditions. From Lemma 3.2 it follows that there are unique values ξi ∈ (ξℵ + ε1, ξ0 − ε1) such that y0(ξj) = zj ∀j = 1, . . . ,ℵ. Evidently, ξℵ+1 < · · · < ξj+1 < ξj < · · · < ξ0. We denote Iε := [ξℵ+1 + ε, ξ0 − ε] \ ℵ⋃ j=1 (ξj − ε, ξj + ε) = ℵ+1⋃ j=1 Ijε , Ijε := [ξj + ε, ξj−1 − ε]. (4.2) From the considerations above it is straightforward to get the following Lemma 4.2. For any arbitrary small positive ε < ε1 one can choose δ > 0 such that for all ξ ∈ Iε the following inequalities are valid inf zj∈σd inf z∈Tδ,j |Re g(z, ξ)| − |Φ(z, ξ)− Φ(zj , ξ)| > C(ε) > 0. (4.3) 20 I. Egorova and J. Michor Note that the infimum in (4.3) is taken along the circles around all points of the discrete spectrum σd. The VDO region{ (n, t) ∈ Z× R+ : n t ∈ Iε } consists of ℵ+ 1 nonintersecting sectors separated by arbitrary small sectors as depicted in Fi- gure 2. If σgap d = ∅, then ℵ = 0 and the VDO region is the simply connected region{ (n, t) ∈ Z× R+ : n t ∈ [ξℵ+1 + ε, ξ0 − ε] } . For each ξ ∈ Iε we divide the eigenvalues based on their relative location with respect to the point µ0(ξ) (respectively, y0(ξ)) and introduce the Blaschke product Π(z) = Π(z, ξ) = ∏ y0(ξ)<zk<0 |zk| z − z−1 k z − zk . (4.4) Note that Π ( z−1 ) = Π−1(z), Π(0) > 0. (4.5) In fact, Π(z, ξ) = Πj(z) = ∏ zj<zk<0 |zk| z − z−1 k z − zk for ξ ∈ Ijε . (4.6) Set also (cf. (2.12)) E(z) = E(z, ξ) =   1 z−zj zjγje 2tΦ(zj) −zjγje2tΦ(zj) z−zj 1  , z ∈ Dδ,j , zj ∈ (y0, 0), σ1E ( z−1 ) σ1, z ∈ D∗δ,j , zj ∈ (y0, 0), I, z ∈ C \ ⋃ zj∈(y0,0) Dδ,j ∪ D∗δ,j . (4.7) The matrix E is not an identity matrix only in small vicinities of those zj which lie in the domain where Re g(z, ξ) > 0. From properties (a), (e) of Lemma 4.1, property (4.5) and (4.7) it follows that if a vector m satisfies (2.17) and (2.18), so does the vector m(1)(z) = m(z)E(z) [ Π(z)et(Φ(z)−g(z))]−σ3 , z ∈ C \ (Σδ ∪ Σ∗δ ∪ J ∪ J∗), (4.8) where σ3 is the third Pauli matrix. A straightforward computation using Theorem 2.3 and Lem- ma 4.1 shows that if m(z) satisfies (2.21), (2.22), then m(1)(z) given by (4.8) solves the jump problem m (1) + (z) = m (1) − (z)v(1)(z), z ∈ Σδ ∪ Σ∗δ ∪ J ∪ J∗, How Discrete Spectrum and Resonances Influence the Asymptotics 21 where v(1)(z) =  ( et(g+(z)−g−(z)) 0 Π−2(z)χ(z) e−t(g+(z)−g−(z)) ) , z ∈ I,( et(g+(z)−g−(z)) 0 0 e−t(g+(z)−g−(z)) ) , z ∈ J,( 1 0 Π−2(z)R(z)e2tg(z) 1 ) , z ∈ Cδ, Aj(z), z ∈ Tδ,j , zj /∈ (y0, 0), Bj(z), z ∈ Tδ,j , zj ∈ (y0, 0), σ1(v(1)(z−1))σ1, z ∈ Σ∗δ ∪ J∗. Here Aj(z) = Aj(z, ξ) = ( 1 0 γjzj Π2(z)(z−zj)e2tHj(z) 1 ) , Bj(z) = Bj(z, ξ) = ( 1 0 Π2(z)(z−zj) γjzj e−2tHj(z) 1 ) , and Hj(z) = Hj(z, ξ) = Φ(zj)− Φ(z) + g(z, ξ). With our choice of the VDO region we evidently have Lemma 4.3. Uniformly with respect to ξ ∈ Iε sup zj∈σd sup z∈Tδ,j (‖Aj(z)− I‖+ ‖Bj(z)− I‖) ≤ C1(ε)e−C(ε)t, (4.9) sup z∈Cδ∪C∗δ ∥∥v(1)(z)− I ∥∥ ≤ C2(ε)e−C(ε)t. (4.10) Here ‖ · ‖ is a norm of 2× 2 matrices. Remark 4.4. The function m(1)(z) inherits the singularities of m(z) described in Theorem 2.3. Recall that these singularities essentially depend on the presence or absence of resonances at points (2.20). In the next step we apply the lense mechanism around I and I∗, which will at the same time weaken these singularities. We will use ` = −1, 0, 1 to indicate singularities as follows. In C \ [q−1, q] introduce a function Q(z) such that Q4(z) =  (z−q)(z−q1) (zq1−1)(zq−1) , q nonresonant, q1 nonresonant (` = −1), (zq1−1)(zq−1) (z−q)(z−q1) , q resonant, q1 resonant (` = 1), (zq1−1)(z−q) (zq−1)(z−q1) , q nonresonant, q1 resonant (` = 0), (zq−1)(z−q1) (zq1−1)(z−q) , q resonant, q1 nonresonant (` = 0), (4.11) with the branch of the forth root defined by the condition Q(1) = 1. Evidently, Q(z−1) = Q−1(z) and Q(0) > 0. It has jumps on I ∪ I∗ for ` = 0 and on I ∪ I∗ ∪ J ∪ J∗ for ` = ±1. The function Ω(z, s) = 1 2s s+ z s− z 22 I. Egorova and J. Michor can be considered as the Cauchy kernel for symmetric contours, because Ω(z, s) = 1 z−s(1 + o(1)) as z → s, and Ω ( z, s−1 ) d ( s−1 ) = Ω ( z−1, s ) ds. Using this kernel allows us to preserve the symmetry condition. Set P(z) = √ (z − q)(z − q1) ( z − q−1 1 )( z − q−1 ) z−2, z ∈ C \ (I ∪ I∗). (4.12) This function satisfies the symmetries P(z−1) = P(z) for z ∈ C \ (I ∪ I∗) and P−(z) = −P+(z) for z ∈ I ∪ I∗. Define S(z) = 1 2πi ∫ q−1 q Ω(z, s)f(s) ds, f(s) :=  log(Π−2(s)Q−2(z)|χ(s)|) P+(s) , s ∈ I, i∆̃ P(s) , s ∈ J, f ( s−1 ) , s ∈ [ q−1,−1 ] , where ∆̃ = ∆̃(ξ) = −i ∫ I log ( Q−2(s)Π−2(s)|χ(s)| ) P+(s) ds s (∫ J ds sP(s) )−1 . It is straightforward to verify that S(z) solves the scalar RH problem S+(z) = S−(z) + f(z), z ∈ [ q, q−1 ] , S(z−1) = −S(z), z ∈ C \ [ q, q−1 ] , S(z) = O(z), z → 0. The above considerations imply that the function F(z) = eP(z)S(z), z ∈ C \ [ q−1, q ] , (4.13) is the unique solution of the following RHP with jump along I ∪ I∗ ∪ J ∪ J∗, (i) F+(z)F−(z) = Π−2(z)|χ(z)|Q−2(z) for z ∈ I, (ii) F+(z) = F−(z)ei∆̃ for z ∈ J , (iii) F(z−1) = F−1(z) for z ∈ C \ [ q−1, q ] , (iv) F(0) > 0, F(1) = 1. Since χ(s)Q−2(s) 6= 0 as s ∈ I ∪ I∗ and it is a continuous function on I ∪ I∗, then F(z) also has nonzero finite limiting values as z → q̃ ∈ J (cf. [21]). It is straightforward to obtain Lemma 4.5. The function F (z) := F(z)Q(z), defined for z ∈ C \ [ q−1, q ] by (4.11)–(4.13), solves the following RHP (i) F+(z)F−(z) = Π−2(z)|χ(z)| for z ∈ I, (ii) F+(z) = F−(z)ei∆ for z ∈ J , (iii) F ( z−1 ) = F−1(z) for z ∈ C \ [ q−1, q ] , (iv) F (0) > 0, F (1) = 1, How Discrete Spectrum and Resonances Influence the Asymptotics 23 where ∆ = ∆(ξ) = −i ∫ I log ( Q−2(s)Π−2(s, ξ)|χ(s)| ) P+(s) ds s (∫ J ds sP(s) )−1 + `π. (4.14) In a vicinity of q̃ ∈ {q, q1} we have F (z) = C(z− q̃)1/4(1+o(1)) if q̃ is a nonresonant point, and F (z) = C(z − q̃)−1/4(1 + o(1)) if q̃ is a resonant point. The jumps of F along the contours I∗ and J∗ as well as its behavior at q−1, q−1 1 are uniquely defined by the symmetry (iii). Note that the only dependence of ∆(ξ) on ξ is due to the Blaschke product (4.4) depending on ξ. It means that ∆(ξ) has constant values on every interval Ijε . By (4.6) we obtain that ∆(ξ) = ∆j for ξ ∈ Ijε , j = 1, . . . ,ℵ+ 1, (4.15) where ∆j = −i ∫ I log ( Q−2(s)Π−2 j (s)|χ(s)| ) P+(s) ds s (∫ J ds sP(s) )−1 + `π. (4.16) Given these preparations, we can implement the next deformation step. Step 2. Introduce two symmetric contours Lδ and L∗δ = { z : z−1 ∈ Lδ } surrounding I and I∗ counterclockwise at a small distance such that min zj∈σd dist ( zj ,Ωδ ) � δ and dist ( y0(ξ),Ωδ ) � δ ∀ξ ∈ Iε, (4.17) where Ωδ and Ω∗δ are the enclosed regions so that Lδ = ∂Ωδ and L∗δ = ∂Ω∗δ , see Figure 5. Condition (4.17) ensures that Lδ is away from the level line Re g(., ξ) = 0 and from any point of the discrete spectrum. 0−1q−1 1q−1 T∗δ,j Cδ C∗δ I∗ Ω∗δ IJ Ωδ L∗δ LδJ∗ Figure 5. Contour deformation of Step 2. Let X(z), z ∈ Ωδ, be the continuation of χ(z) as described in Remark 2.4. Set GF (z) = ( F−1(z) −Π2(z)F (z) X(z) e−2tg(z) 0 F (z) ) , z ∈ Ωδ. Define m(2)(z) by m(2)(z) =  m(1)(z)GF (z), z ∈ Ωδ, m(2) ( z−1 ) σ1, z ∈ Ω∗δ , m(1)(z)(F (z))−σ3 , z ∈ C \ (Ωδ ∪ Ω∗δ). 24 I. Egorova and J. Michor Theorem 4.6. For every ξ ∈ Iε, the vector function m(2)(z) = m(2)(z, ξ) is the unique solution of the following RHP: to find a holomorphic function in the domain C \ (Σδ ∪ Σ∗δ ∪ Lδ ∪ L∗δ ∪ J ∪ J∗) which has continuous limits on the sides of the contour Σδ ∪Σ∗δ ∪Lδ ∪L∗δ ∪J ∪J∗ except possibly at points J (2.20) and satisfies: � the jump condition m (2) + (z, n, t) = m (2) − (z, n, t)v(2)(z, n, t), v(2)(z) =  vmod(z), z ∈ I ∪ I∗ ∪ J ∪ J∗,( 1 Π2(z)F 2(z) X(z) e−2tg(z) 0 1 ) , z ∈ Lδ, [F (z)]−σ3v(1)(z)[F (z)]σ3 , z ∈ Σδ \ I, σ1 ( v(2) ( z−1 )) σ1, z ∈ Σ∗δ ∪ J∗ ∪ L∗δ , (4.18) where Σδ is defined by (2.13) and vmod(z) =  iσ1, z ∈ I, e−(nΛ+tU+i∆)σ3 , z ∈ J σ1v mod ( z−1 ) σ1, z ∈ I∗ ∪ J∗, (4.19) � the symmetry condition m(2) ( z−1 ) = m(2)(z)σ1, � the normalization condition m (2) 1 (0) ·m(2) 2 (0) = 1, m (2) 1 (0) > 0, � at points of the set (2.20), m(2)(z) has at most a fourth root singularity, m(2)(z) = O(z − κ)−1/4, as z → κ ∈ J . Proof. The proof of this theorem is completely analogous to the proof of [11, Theorem 3.6] except for a small contour Iδ = [q1, q1 − δ] = Ωδ ∩ [q1,−1] ⊂ J , where the jump matrix v(2)(z) = [ GF−(z) ]−1 v(1)(z)GF+(z) should be evaluated. From Lemmas 4.1(f), and 4.5(ii), it follows that v(2) = ( F− Π2F− X e−2tg− 0 F−1 − )( et(g+−g−) 0 0 et(g−−g+) )( F−1 + −Π2F+ X e−2tg+ 0 F+ ) = ( F− F+ et(g+−g−) Π2e−t(g++g−) X (F+F− − F+F−) 0 F+ F− et(g−−g+) ) = ( e−nΛ−tU−i∆ 0 0 enΛ+tU+i∆ ) . On the symmetric contour Iδ,∗ (oriented from left to right) we get v(2) = vmod by the sym- metry. � Remark 4.7. According to (4.16) and (4.19) we see that vmod(z) = vmod j (z, n, t) =  e−(nΛ+tU+i∆j)σ3 , z ∈ J, n t ∈ I j ε , iσ1, z ∈ I, σ1v mod ( z−1 ) σ1, z ∈ I∗ ∪ J∗. (4.20) How Discrete Spectrum and Resonances Influence the Asymptotics 25 Let us label the jump contour which appears in Theorem 4.6 by Kδ = Σδ ∪ Σ∗δ ∪ Lδ ∪ L∗δ ∪ J ∪ J∗. (4.21) We extend the matrix vmod(z) to the whole contour Kδ by defining it as the identity matrix on the remaining part Kδ \ (I ∪ I∗ ∪ J ∪ J∗). Lemma 4.8. Uniformly with respect to ξ ∈ Iε, sup z∈Kδ ∥∥v(2)(z)− vmod(z) ∥∥ ≤ C3(ε)e−C4(ε)t, Ci(ε) > 0. (4.22) Proof. Recall that inf z∈Lδ Re g(z, ξ) > C5(ε) > 0 ∀ξ ∈ Iε. This inequality verifies (4.22) on the contour Lδ ∪ L∗δ . Since the conjugation [F (z)]−σ3v(1)(z)[F (z)]σ3 , z ∈ Σδ \ I, does not impair estimates (4.9), (4.10) and since v(2)(z) = vmod(z) for z ∈ I ∪ I∗ ∪ J ∪ J∗, then (4.22) is straightforward for the remaining part of Kδ. � 5 Solution of the model problem and conclusive analysis In Section 3, Lemma 3.3, we constructed the vector-function m̃(p), which solves the jump problem (3.13) for ∆ given by (3.12). One can treat this result in the following way: let ∆ be an arbitrary real value and let p0 be the unique solution of the Jacobi inversion problem (3.12). Consider this point as the initial Dirichlet divisor and let â(n, t,∆), b̂(n, t,∆) be the finite gap solution associated with this divisor and with the spectrum S. In particular, we can construct m̃(p) associated with ∆ given by (4.14). Being considered on the z-plane, the vector-function m̃(z) = m̃(p(z)) has additional jumps on I ∪ I∗ due to (3.14) and solves the jump problem m̃+(z) = m̃−(z) { σ1, z ∈ I ∪ I∗, vmod(z), z ∈ J ∪ J∗, with vmod(z) given by (4.19) (or by (4.20)) on J ∪ J∗. Introduce the function H(z) = 4 √ (q1 − z) ( q−1 1 − z ) (q − z) ( q−1 − z ) , which satisfies H ( z−1 ) = H(z), z ∈ C \ (I ∪ I∗); H+(z) = iH−(z), z ∈ I ∪ I∗; H(0) = 1. Thus mmod(z) := H(z)m̃(z) is the unique solution of the following Model RH problem. Find a vector-function mmod(z) holomorphic in C \ [ q, q−1 ] , continuous up to the boundary except of points of the set (2.20), which satisfies the jump condition mmod + (z) = mmod − (z)vmod(z), z ∈ I ∪ I∗ ∪ J ∪ J∗, with vmod(z) given by (4.19), and the symmetry and normalization conditions mmod ( z−1 ) = mmod(z)σ1, mmod 1 (0) = [ mmod 2 (0) ]−1 > 0. 26 I. Egorova and J. Michor At points of the set (2.20) it has a fourth root singularity mmod(z) = O(z − κ)−1/4, as z → κ ∈ J . Uniqueness of the solution of such a problem was established in [12]. The dependence of mmod(z) on n, t and ∆ ( n t ) is due to the jump exp (( −nΛ−tU+i∆ ( n t )) σ3 ) . For large n and t, if n t ∈ I j ε , then n+1 t ∈ I j ε . Recall that ∆(ξ) has constant values ∆j on Ijε (cf. (4.16)). Therefore, mmod(z, n, t, j) := mmod(z, n, t,∆j) and mmod(z, n+ 1, t, j) := mmod(z, n+ 1, t,∆j) are well defined for n t ∈ I j ε . Here mmod(z, n + 1, t, j) is the solution of the jump problem with the jump exp (−((n+ 1)Λ + tU + i∆j)σ3) on J and associated jumps on J∗ ∪ I ∪ I∗. Since H2(z) = ( 1− z 2q1 − zq1 2 +O ( z2 ))( 1 + z 2q + zq 2 +O ( z2 )) = 1− 2z(2a) +O ( z2 ) and λ = 1 2z (1 + o(1)), then Theorem 3.6 and (4.15) imply Lemma 5.1. For all n→∞, t→∞ and n t ∈ I j ε , lim z→0 mmod 1 (z, n, t, j) mmod 1 (z, n+ 1, t, j) = â(n, t,∆j)ã −1, lim z→0 1 2z ( mmod 1 (z, n, t, j)mmod 2 (z, n, t, j)− 1 ) = b̂(n, t,∆j). (5.1) Here the phase ∆j is defined by (4.16), (4.6) and ã = CapS. Our next task is to prove the following approximation Theorem 5.2. For all n → ∞, t → ∞ such that n t ∈ I j ε , the following asymptotic holds as z → 0 m(2)(z, n, t) = mmod(z, n, t, j) +O ( e−C(ε)t ) (1 +O(z)), C(ε) > 0. (5.2) Moreover, m1(z)m2(z) = mmod 1 (z, n, t, j)mmod 2 (z, n, t, j) + βj1(ξ, t) + βj2(ξ, t)z + βj3(ξ, t)O ( z2 ) , z → 0, (5.3) where βjk(ξ, t) = O ( e−C(ε)t ) , k = 1, 2, 3, uniformly with respect to ξ ∈ Ijε . The proof of Theorem 5.2 essentially repeats the arguments used in [11, Section 7]. But since this theorem provides the justification of our asymptotic analysis, we will briefly describe the key points of the final analysis as applied to our case. According to the standard approach (cf. [3, 16]), to perform the conclusive analysis we first have to evaluate in C the “error vector” merr(z, n, t, j) := m(2)(z, n, t) [ Mmod j (z, n, t) ]−1 , (5.4) where Mmod j (z, n, t) is a matrix solution of the model RHP. Recall that by Remark 4.7, the model jump problem Mmod j,+ (z, n, t) = Mmod j,− (z, n, t)vmod j (z, n, t), z ∈ I ∪ I∗ ∪ J ∪ J∗, n t ∈ Ijε , (5.5) How Discrete Spectrum and Resonances Influence the Asymptotics 27 has a piecewise constant (with respect to ξ) jump matrix vmod(z) given by (4.20). The choice of such a matrix solution is not unique. However, a proper solution Mj(z) := Mmod j (z, n, t) should be invertible and satisfy at least the symmetry Mj ( z−1 ) = σ1Mj(z)σ1, because the vector merr(z, n, t, j) should fulfill the standard symmetry property. Moreover, the singularities of merr should be removable outside the jump contour and should not be more than L2-integrable on the contour. By construction (5.4), merr does not have a jump on I ∪ I∗ ∪ J ∪ J∗. Since m(2)(z) has singularities of order O(z−q̃)−1/4 at q̃ ∈ J (2.20), the matrix [Mj(z)] −1 should have singularities of order not exceeding o(z − q̃)−3/4 at q̃ ∈ J and should be less than poles at its other singular points outside the contour. As it is shown in [13] for the KdV shock wave (and the same is true for the Toda shock case) such an invertible solution with weak singularities does not exist for certain arbitrary large t at any direction ξ ∈ Iδ. Instead, we choose a matrix solution with poles at points 1,−1, such that its determinant is equal to 1 and merr has removable singularities at these points. To explain the construction of our solution in more detail, recall that the quasimomen- tum (3.1) has the jump on J ∪ J∗ defined by (3.10). Introduce the function G(z) := exp (∫ p(z) b−2a ω∞+,∞− ) , z ∈ C \ [ q, q−1 ] . Recall that for z ∈ D we have p(z) = ( z+z−1 2 ,+ ) ∈ M (upper sheet), and for z ∈ D∗ we have p(z) = ( z+z−1 2 ,− ) . Then G(z) has the following properties [11, Section 5]: � G(z) is holomorphic on C \ [ q, q−1 ] and satisfies G ( z−1 ) = G−1(z). � Its jumps are given by G+(z) = G−(z)e−Λ, z ∈ J, G+(z) = G−(z)eΛ, z ∈ J∗, G±(z) = [ G± ( z−1 )]−1 , z ∈ I ∪ I∗. � The following asymptotic expansion is valid, G(z) = − ã 2z ( 1 + 2b̃z +O ( z2 )) , G(z−1) = −2z ã ( 1− 2b̃z +O ( z2 )) , where ã = CapS and b̃ ∈ R. Lemma 5.3. Introduce the following vector function holomorphic in C \ ([q, q−1] ∪ {0}), m#(z, n, t, j) = mmod(z, n+ 1, t, j)[G(z)]−σ3 . Then m#(z) solves the jump problem m# +(z, n, t, j) = m# −(z, n, t, j)vmod j (z, n, t), z ∈ I ∪ I∗ ∪ J ∪ J∗, n t ∈ Ijε , where vmod j (z, n, t) is given by (4.20). The vector m#(z) satisfies the symmetry condition m# 2 ( z−1 ) = m# 1 (z). The normalization condition is not fulfilled, instead we have m# 1 (z, n, t, j) = −2z ã mmod 1 (0, n+ 1, t, j)(1 +O(z)), m# 2 (z, n, t, j) = − ã 2z mmod 1 (∞, n+ 1, t, j)(1 +O(z)), as z → 0. 28 I. Egorova and J. Michor Proof. The proof follows immediately from the properties of G and mmod above. � The function ρ(z) = ρ(z, n, t, j) = mmod 2 (0, n, t, j)mmod 2 (∞, n+ 1, t, j) 2ã ( z−1 − z ) is defined for all z 6= ±1 and is odd, ρ(z) = −ρ ( z−1 ) . The function ρ does not have jumps, there- fore the vector ρ(z)m#(z) solves the same jump problem with (4.20). However, it is bounded as z → 0, z →∞ and has simple poles at 1 and −1 instead. In conclusion, the vector Ψ(z, n, t, j) = 1 2 mmod(z, n, t, j) + ρ(z)m#(z, n, t, j) solves our vector model RHP, and the same is true for Ψ ( z−1, n, t, j ) σ1. From here on, we fix the parameters n, t, j and omit them to shorten notations when necessary. In particular, the symmetry conditions for mmod and m# and oddness of ρ imply for the components of Ψ that Ψ1(z) = 1 2 mmod 1 (z) + ρ(z)m# 1 (z), Ψ2(z) = 1 2 mmod 2 (z) + ρ(z)m# 2 (z), Ψ1 ( z−1 ) = 1 2 mmod 2 (z)− ρ(z)m# 2 (z), Ψ2 ( z−1 ) = 1 2 mmod 1 (z)− ρ(z)m# 1 (z). (5.6) Lemma 5.4 ([11, Lemma 5.2]). (i) The matrix Mmod(z) = Mmod j (z, n, t) Mmod(z) = ( Ψ1(z) Ψ2(z) Ψ2(z−1) Ψ1(z−1) ) , z ∈ C \ [ q−1, q ] , is a meromorphic matrix solution for the model jump problem Mmod + (z) = Mmod − (z)vmod(z), z ∈ I ∪ I∗ ∪ J ∪ J∗, with vmod(z) given by (4.19). It has simple poles at z = ±1. (ii) Mmod(z) satisfies the symmetry Mmod(z−1) = σ1M mod(z)σ1. (5.7) (iii) The following equality is valid, mmod(z, n, t) = (1, 1)Mmod(z, n, t). (5.8) (iv) The determinant of Mmod(z) is a constant with respect to n, t, j, z, detMmod(z) = 1, z ∈ C. (5.9) Using this lemma, we can establish the properties of the error vector function (5.4), which we denote here by merr(z). We recall the definition of Kδ from (4.21). Theorem 5.5. The vector function merr(z) is holomorphic in C \ Ξδ, where Ξδ := Lδ ∪ L∗δ ∪ Cδ ∪ C∗δ ∪ ⋃ zk∈σd ( Tδ,k ∪ T∗δ,k ) = Kδ \ (I ∪ I∗ ∪ J ∪ J∗), and satisfies the following properties: How Discrete Spectrum and Resonances Influence the Asymptotics 29 (i) merr(z) has removable singularities at 0, 1, −1, ∞ and on J . (ii) It solves the jump problem merr + (z) = merr − (z)(I +W (z)), z ∈ Ξδ, (5.10) where W (z) = Mmod(z) ( v(2)(z)− I )[ Mmod(z) ]−1 . (5.11) (iii) It satisfies the symmetry condition merr(z) = merr ( z−1 ) σ1, z ∈ C \ Ξδ. (5.12) In particular, merr − ( z−1 ) = merr − (z)σ1, z ∈ Ξδ, (5.13) merr 2 (0) = merr 1 (∞) = 1 2 ( m (2) 1 (0) mmod 1 (0) + mmod 1 (0)) m (2) 1 (0) ) := τ > 0. (5.14) Here mmod 1 (0) = mmod 1 (0, n, t, j) and m(2)(0) = m(2)(0, n, t). Proof. The absence of singularities at the points 0, 1, −1,∞ and on the set J was proven in [11, Lemmas 5.4 and 5.5]. The jump (5.10) and the symmetry property (5.12) follow from (4.18), (4.19), (5.5), (5.7) and (5.9). Property (5.13) holds due to the mutual orientation of the sym- metric parts of the contour Ξδ. Last, (5.14) follows from (5.4), (5.6), (5.9) and the definition of ρ which implies lim z→0 ρ(z)m# 2 (z) = − 1 mmod 1 (0) , lim z→0 ρ(z)m# 1 (z) = 0. � Now we are ready to prove Theorem 5.2. We follow the well-known approach via singular integral equations (see, e.g., [9], [16, Chapter 4], [19]). A peculiarity of this approach applied to the Toda equation is generated by the type of normalization condition of the vector RHP and the symmetry condition. In particular, if we want to preserve the symmetry condition (5.12) in the Cauchy-type formula for merr(z), we should use a matrix Cauchy kernel (cf. [19, equation (B.8)], Ω̂(s, z) = ( 1 s−z 0 0 1 s−z − 1 s ) ds, s ∈ Ξδ, z /∈ Ξδ. Since [19, equation (B.9)] Ω̂ ( s, z−1 ) = σ1Ω̂ ( s−1, z ) σ1, and W (s−1) = σ1W (s)σ1, s ∈ Ξδ, this implies with (5.13) and the orientation of Ξδ that the symmetry property holds:∫ Ξδ merr − (s)W (s)Ω̂(s, z) = ∫ Ξδ merr − (s)W (s)Ω̂ ( s, z−1 ) σ1. Note that the 1, 1-entry of the Cauchy kernel Ω̂(s, z) has a zero at z = ∞ while the 2, 2-entry has a zero at z = 0. From (5.10) and (5.14) it follows that merr(z) = ( merr 1 (∞),merr 2 (0) ) + 1 2πi ∫ Ξδ merr − (s)W (s)Ω̂(s, z) = τ(1, 1) + 1 2πi ∫ Ξδ merr − (s)W (s)Ω̂(s, z). 30 I. Egorova and J. Michor Recall that according to Lemma 4.8, ‖W (z)‖ = O ( e−C(δ)t ) , uniformly with respect to z ∈ Ξδ and n, t, j : n t ∈ Ijε . In turn, it implies the following Lemma 5.6. Uniformly with respect to n t ∈ Iε,∥∥zkW (z) ∥∥ Lp(Ξδ) = O ( e−C(δ)t ) , p ∈ [1,∞], k = 0, 1. (5.15) Now we are ready to apply the technique of singular integral equations. Let C denote the Cauchy operator associated with Ξδ, (Ch)(z) = 1 2πi ∫ Ξδ h(s)Ω̂(s, z), s ∈ C \ Ξδ, where h = ( h1 h2 ) ∈ L2(Ξδ) and satisfies the symmetry h(s) = h ( s−1 ) σ1. Let (C+h)(z) and (C−h)(z) be the non-tangential limiting values of (Ch)(z) from the left and right sides of Ξδ, respectively. As usual, we introduce the operator CW : L2(Ξδ) ∩ L∞(Ξδ) → L2(Ξδ) by CWh = C−(hW ), where W is the error matrix (5.11). Then ‖CW ‖ = ‖CW ‖L2(Ξδ)→L2(Ξδ) ≤ C‖W‖L∞(Ξδ) = O ( e−C(ε)t ) , as well as∥∥(I− CW )−1 ∥∥ = ∥∥(I− CW )−1 ∥∥ L2(Ξδ)→L2(Ξδ) ≤ 1 1−O ( e−C(ε)t ) (5.16) for sufficiently large t. Consequently, for t� 1, on Ξδ we define a vector function µ(s) = (τ, τ) + (I− CW )−1CW ( (τ, τ) ) (s), with τ given by (5.14). Then by (5.15) and (5.16) ‖µ(s)− (τ, τ)‖L2(Ξδ) ≤ ∥∥(I− CW )−1 ∥∥‖C−‖‖W‖L∞(Ξδ) = O ( e−C(ε)t ) . (5.17) With the help of µ, the vector function merr(z) can be represented as merr(z) = (τ, τ) + 1 2πi ∫ Ξδ µ(s)W (s)Ω̂(s, z), and by virtue of (5.17) and Lemma 5.6 we obtain as z → 0 merr(z) = (τ, τ) + 1 2πi ∫ Ξδ (τ, τ)W (s)Ω̂(s, z) + E(z). (5.18) Here E(z) is a vector function holomorphic in a vicinity of z = 0 which admits the estimate ‖E(z)‖ ≤ ‖W‖L2(Ξδ)‖µ(s)− (τ, τ)‖L2(Ξδ)(1 +O(z)) = O ( e−C(ε)t ) (1 +O(z)), and O(z) is uniformly bounded for n t ∈ I j ε . From (5.18) and (5.8) we get m(2)(z) = merr(z)Mmod(z) = τmmod(z) + τO ( e−C(ε)t ) E1(z), How Discrete Spectrum and Resonances Influence the Asymptotics 31 where E1(z) is a holomorphic vector function in a vicinity of z = 0, uniformly bounded with respect to n, t, j as n t ∈ I j ε . The normalization conditions for m(2) and mmod imply that τ2 ( 1 +O ( e−C(ε)t )) = 1, that is, τ = 1 +O ( e−C(ε)t ) . With (5.14) this yields (5.2) and at the same time (5.3), which proves Theorem 5.2. To finish the proof of Theorem 1.1 recall that in a vicinity of z = 0 the initial vector- function (2.14) and the transformed function m(2)(z) are connected by m(2)(z, n, t) = m(z, n, t) [ Π(z, ξ)et(Φ(z,ξ)−g(z,ξ))F (z, ξ) ]−σ3 , ξ = n t . (5.19) The results of Theorem 5.2 and Lemmas 2.2 and 5.1 imply b(n, t) = b̂(n, t,∆j) +O ( e−C(ε)t ) , n t ∈ Ijε . On the other hand, from (2.16), (5.19) and Lemma 4.1(e), it follows 2a(n, t) = m1(0, n, t) m1(0, n+ 1, t) = m (2) 1 (0, n, t) m (2) 1 (0, n+ 1, t) et(K(nt )−K(n+1 t )), because Π(z) and F (z) are the same for n and n + 1. By (4.1), K(ξ) = − log(2ã)(ξ − const), that is, t ( K ( n t ) −K ( n+ 1 t )) = log(2ã). Using (5.1) we finally get 2a(n, t) = mmod 1 (0, n, t) mmod 1 (0, n+ 1, t) e2ã +O ( e−C(ε)t ) = 2â(n, t,∆j) +O ( e−C(ε)t ) . This finishes the proof of our main result, Theorem 1.1. Acknowledgements This research was supported by the Austrian Science Fund (FWF) under Grant No. P31651. We thank the referees for their careful reading and their recommendations. References [1] Andreiev K., Egorova I., Lange T.L., Teschl G., Rarefaction waves of the Korteweg–de Vries equation via nonlinear steepest descent, J. Differential Equations 261 (2016), 5371–5410, arXiv:1602.02427. [2] Bilman D., Trogdon T., Numerical inverse scattering for the Toda lattice, Comm. Math. Phys. 352 (2017), 805–879, arXiv:1508.01788. [3] Bleher P.M., Lectures on random matrix models: the Riemann–Hilbert approach, in Random Matrices, Random Processes and Integrable Systems, CRM Ser. Math. Phys., Springer, New York, 2011, 251–349, arXiv:0801.1858. [4] Bloch A.M., Kodama Y., Dispersive regularization of the Whitham equation for the Toda lattice, SIAM J. Appl. Math. 52 (1992), 909–928. [5] Bloch A.M., Kodama Y., The Whitham equation and shocks in the Toda lattice, in Singular Limits of Dispersive Waves (Lyon, 1991), NATO Adv. Sci. Inst. Ser. B Phys., Vol. 320, Plenum, New York, 1994, 1–19. https://doi.org/10.1016/j.jde.2016.08.009 https://arxiv.org/abs/1602.02427 https://doi.org/10.1007/s00220-016-2819-0 https://arxiv.org/abs/1508.01788 https://doi.org/10.1007/978-1-4419-9514-8_4 https://arxiv.org/abs/0801.1858 https://doi.org/10.1137/0152052 https://doi.org/10.1137/0152052 32 I. Egorova and J. Michor [6] Deift P., Kamvissis S., Kriecherbauer T., Zhou X., The Toda rarefaction problem, Comm. Pure Appl. Math. 49 (1996), 35–83. [7] Deift P., Kriecherbauer T., McLaughlin K.T.-R., Venakides S., Zhou X., Uniform asymptotics for poly- nomials orthogonal with respect to varying exponential weights and applications to universality questions in random matrix theory, Comm. Pure Appl. Math. 52 (1999), 1335–1425. [8] Deift P., Venakides S., Zhou X., The collisionless shock region for the long-time behavior of solutions of the KdV equation, Comm. Pure Appl. Math. 47 (1994), 199–206. [9] Deift P., Zhou X., A steepest descent method for oscillatory Riemann–Hilbert problems. Asymptotics for the MKdV equation, Ann. of Math. 137 (1993), 295–368, arXiv:math.AP/9201261. [10] Egorova I., Gladka Z., Kotlyarov V., Teschl G., Long-time asymptotics for the Korteweg–de Vries equation with step-like initial data, Nonlinearity 26 (2013), 1839–1864, arXiv:1210.7434. [11] Egorova I., Michor J., Pryimak A., Teschl G., Long-time asymptotics for Toda shock waves in the modulation region, arXiv:2001.05184. [12] Egorova I., Michor J., Teschl G., Long-time asymptotics for the Toda shock problem: non-overlapping spectra, J. Math. Phys. Anal. Geom. 14 (2018), 406–451, arXiv:1406.0720. [13] Egorova I., Piorkowski M., Teschl G., On vector and matrix Riemann–Hilbert problems for KdV shock waves, arXiv:1907.09792. [14] Girotti M., Grava T., Jenkins R., McLaughlin K.D.T.-R., Numerical inverse scattering for the Toda lattice, Comm. Math. Phys., to appear, arXiv:1807.00608. [15] Grunert K., Teschl G., Long-time asymptotics for the Korteweg–de Vries equation via nonlinear steepest descent, Math. Phys. Anal. Geom. 12 (2009), 287–324, arXiv:0807.5041. [16] Its A.R., Large N asymptotics in random matrices: the Riemann–Hilbert approach, in Random Matrices, Random Processes and Integrable Systems, CRM Ser. Math. Phys., Springer, New York, 2011, 351–413. [17] Kamvissis S., On the Toda shock problem, Phys. D 65 (1993), 242–266. [18] Krüger H., Teschl G., Long-time asymptotics for the Toda lattice in the soliton region, Math. Z. 262 (2009), 585–602, arXiv:0711.2793. [19] Krüger H., Teschl G., Long-time asymptotics of the Toda lattice for decaying initial data revisited, Rev. Math. Phys. 21 (2009), 61–109, arXiv:0804.4693. [20] Michor J., Wave phenomena of the Toda lattice with steplike initial data, Phys. Lett. A 380 (2016), 1110– 1116, arXiv:1510.03581. [21] Muskhelishvili N.I., Singular integral equations, Wolters-Noordhoff Publishing, Groningen, 1972. [22] Piorkowski M., Parametrix problem for the Korteweg–de Vries equation with steplike initial data, arXiv:1908.11340. [23] Teschl G., Jacobi operators and completely integrable nonlinear lattices, Mathematical Surveys and Mono- graphs, Vol. 72, Amer. Math. Soc., Providence, RI, 2000. [24] Toda M., Theory of nonlinear lattices, 2nd ed., Springer Series in Solid-State Sciences, Vol. 20, Springer- Verlag, Berlin, 1989. [25] Venakides S., Deift P., Oba R., The Toda shock problem, Comm. Pure Appl. Math. 44 (1991), 1171–1242. https://doi.org/10.1002/(SICI)1097-0312(199601)49:1<35::AID-CPA2>3.0.CO;2-8 https://doi.org/10.1002/(SICI)1097-0312(199911)52:11<1335::AID-CPA1>3.0.CO;2-1 https://doi.org/10.1002/cpa.3160470204 https://doi.org/10.2307/2946540 https://arxiv.org/abs/math.AP/9201261 https://doi.org/10.1088/0951-7715/26/7/1839 https://arxiv.org/abs/1210.7434 https://arxiv.org/abs/2001.05184 https://doi.org/10.15407/mag14.04.406 https://arxiv.org/abs/1406.0720 https://arxiv.org/abs/1907.09792 https://doi.org/10.1007/s00220-021-03942-1 https://arxiv.org/abs/1807.00608 https://doi.org/10.1007/s11040-009-9062-2 https://arxiv.org/abs/0807.5041 https://doi.org/10.1007/978-1-4419-9514-8_5 https://doi.org/10.1016/0167-2789(93)90161-S https://doi.org/10.1007/s00209-008-0391-9 https://arxiv.org/abs/0711.2793 https://doi.org/10.1142/S0129055X0900358X https://doi.org/10.1142/S0129055X0900358X https://arxiv.org/abs/0804.4693 https://doi.org/10.1016/j.physleta.2016.01.033 https://arxiv.org/abs/1510.03581 https://arxiv.org/abs/1908.11340 https://doi.org/10.1090/surv/072 https://doi.org/10.1007/978-3-642-83219-2 https://doi.org/10.1007/978-3-642-83219-2 https://doi.org/10.1002/cpa.3160440823 1 Introduction 2 Notations and statement of the initial holomorphic RHP 3 g-function as an Abel integral and its connection with the Baker–Akhiezer function 4 Reduction of the initial RHP to the model RHP 5 Solution of the model problem and conclusive analysis References
id nasplib_isofts_kiev_ua-123456789-211304
institution Digital Library of Periodicals of National Academy of Sciences of Ukraine
issn 1815-0659
language English
last_indexed 2026-03-19T02:46:08Z
publishDate 2021
publisher Інститут математики НАН України
record_format dspace
spelling Egorova, Iryna
Michor, Johanna
2025-12-29T11:06:38Z
2021
How Discrete Spectrum and Resonances Influence the Asymptotics of the Toda Shock Wave. Iryna Egorova and Johanna Michor. SIGMA 17 (2021), 045, 32 pages
1815-0659
2020 Mathematics Subject Classification: 37K40; 35Q53; 37K45; 35Q15
arXiv:2012.12371
https://nasplib.isofts.kiev.ua/handle/123456789/211304
https://doi.org/10.3842/SIGMA.2021.045
We rigorously derive the long-time asymptotics of the Toda shock wave in a middle region where the solution is asymptotically a finite gap. In particular, we describe the influence of the discrete spectrum in the spectral gap on the shift of the phase in the theta-function representation for this solution. We also study the effect of possible resonances at the endpoints of the gap on this phase. This paper is a continuation of research started in [arXiv:2001.05184].
This research was supported by the Austrian Science Fund (FWF) under Grant No. P31651. We thank the referees for their careful reading and their recommendations.
en
Інститут математики НАН України
Symmetry, Integrability and Geometry: Methods and Applications
How Discrete Spectrum and Resonances Influence the Asymptotics of the Toda Shock Wave
Article
published earlier
spellingShingle How Discrete Spectrum and Resonances Influence the Asymptotics of the Toda Shock Wave
Egorova, Iryna
Michor, Johanna
title How Discrete Spectrum and Resonances Influence the Asymptotics of the Toda Shock Wave
title_full How Discrete Spectrum and Resonances Influence the Asymptotics of the Toda Shock Wave
title_fullStr How Discrete Spectrum and Resonances Influence the Asymptotics of the Toda Shock Wave
title_full_unstemmed How Discrete Spectrum and Resonances Influence the Asymptotics of the Toda Shock Wave
title_short How Discrete Spectrum and Resonances Influence the Asymptotics of the Toda Shock Wave
title_sort how discrete spectrum and resonances influence the asymptotics of the toda shock wave
url https://nasplib.isofts.kiev.ua/handle/123456789/211304
work_keys_str_mv AT egorovairyna howdiscretespectrumandresonancesinfluencetheasymptoticsofthetodashockwave
AT michorjohanna howdiscretespectrumandresonancesinfluencetheasymptoticsofthetodashockwave