Asymptotic Lower Bound on the Spatial Analyticity Radius for Solutions of the Periodic Fifth Order KdV–BBM Equation
Abstract
In this work, consideration is given to the initial value problem associated with the periodic fifth-order KdV–BBM equation. It is shown that the uniform radius of spatial analyticity σ(t) of solution at time t is bounded from below by ct−2/3 (for some c > 0), given initial data η0 that is analytic on the circle and has a uniform radius of spatial analyticity σ0. The proof of our main theorems is based on a contraction mapping argument, a method of approximate conservation law in a modified Gevrey spaces, Hölder’s inequality, Sobolev algebra, Cauchy–Schwartz inequality, and Sobolev embedding.
1. Introduction
The IVPs for some dispersive equations have plenty of significant issues under the periodic setting compared to nonperiodic IVPs. The problem that makes a difference from the nonperiodic IVP is the presence of nontrivial resonance. In what follows, the fifth-order KdV–BBM equation (1) describes the unidirectional propagation of water waves, and this was introduced by Bona et al. [1] by using the second-order approximation in the two-way model, the so-called abcd-system derived in [4, 5].
Local well-posedness of the IVP (1) with data in the Sobolev spaces for s ≥ 1 was established by Carvajal et al. [2]. Moreover, for γ1, δ1 > 0 and γ = 7/48, the authors used the conserved energy quantity to prove global well-posedness of (1) for data in , s ≥ 2. Furthermore, the authors used the method of splitting argument to improve the global well-posedness result in for 1 ≤ s < 2. In addition to this, the authors also showed that the IVP (1) is ill-posed in the sense that the flow-map that takes initial data to solution can not be continuous for given data in . In line with this, the decay rate for (1) on the real line has been studied in [3, 6, 7].
It is clear that G0,s = H0,s = Hs, where denotes the L2-based Sobolev space of order s.
The interest in the space is due to (6), and the Paley–Wiener Theorem [11], which states that for σ > 0 and s ≥ 0, a 2π-periodic function f belongs to if and only if f is the restriction to the real line of a function F, which is holomorphic in the strip and satisfies the bound .
Information about the domain of analyticity of a solution to a partial differential equation (PDE) can be used to gain a quantitative understanding of the structure of the equation, and to obtain insight into underlying physical processes. The idea of obtaining spatial analyticity was introduced by Kato and Masuda [12]. There are various studies concerning the properties of spatial analyticity of solutions for a large class of nonlinear partial differential equations. For instance, parabolic equations [13], Navier–Stoke equation [14], cubic Szegő equation [15], periodic generalized KdV equation [16,17], generalized Euler equation [18], second-order analytic nonlinear differential equations [19], semilinear Klein–Gordon equations [20] and the references therein.
Rapid progress has been made lately in obtaining an algebraic decay rate of the radius, i.e., σ(t) ~ t−α for some α ≥ 1, to various nonlinear dispersive PDEs. For example, Himonas et al. in [21] studied the persistence of spatial analyticity for periodic solutions of the dispersion-generalized (g) KdV equation for β ≥ 2. For a class of analytic initial data with a uniform radius of analyticity σ0 > 0, the authors obtained an asymptotic lower bound σ(t) ≥ ct−p on the uniform radius of analyticity σ(t) at time t, as t⟶∞, where p = max{1, 4/β}. Moreover, Tesfahun [22] showed that the uniform radius of spatial analyticity σ(t) of solutions u(t) at time t to the KdV equation cannot decay faster than t−4/3 as t⟶∞ given initial data that is analytic with fixed radius σ0 > 0. Furthermore, in a paper [23], Himonas and Petronilho studied the IVP of the Benjamin–Bona–Mahony (BBM) equation with initial data that are analytic on the torus and have a uniform radius of analyticity σ0 > 0. In their work, they examined the evolution of the radius of spatial analyticity σ(t) of the solution u(t) at any future time t. The authors showed that the size of the radius of spatial analyticity persists for some time and after that, it evolves in such a way that its size at any time t is bounded below by ct−1 for some c > 0. A similar lower bound of spatial analyticity was obtained in [24, 25] for the Schrödinger equation with cubic nonlinearity on both the real line and the circle. The method used in these papers was first introduced by Selberg and Tesfahun [8] in the context of the 1d Dirac–Klein–Gordon equations, which is based on approximate conservation laws in a Gevrey space and Bourgan’s Fourier restriction method.
As a consequence of property (9) and the existing well-posedness theory in (see [2]), we conclude that the IVP (1) with initial data η0 in for all σ0 > 0 and s ≥ 1 have a unique and global in time solution provide that γ1, δ1 > 0 and γ = 7/48.
The main result in this paper is as follows.
Theorem 1. Assume γ1, δ1 > 0. Let s ≥ 1, γ = 7/48, and for σ0 > 0. Then, the global solution η of (1) satisfies
Notation: throughout this paper, we use C to denote generic constant that may vary at each occurrence. For any positive quantities a and b, we use a≲b to indicate an estimate of the form a ≤ Cb. Moreover, we write a ~ b if a≲b and b≲a. Furthermore, we denote by a+ the quantity a + ε for any ε > 0 and by Dα the differential operator given by .
2. Local Well-Posedness in Hσ,s
In this section, we prove the local well-posedness of the IVP (1) in the space Hσ,s for s ≥ 1 and σ > 0, using multilinear estimates in combination with a contraction mapping argument. In fact, we outline the argument in [3] that enables the authors to obtain the local well-posedness result for the IVP associated to the fifth oder KdV–BBM model on the real line in Gσ,s with s ≥ 1, σ > 0.
Lemma 1 (see [23], Lemma 2.1, [2], Proposition 2.1.)Let with σ ≥ 0, and s ≥ 0. Then, we have
Lemma 2. Let τ(D) be the operator given in (6), σ > 0, and s ≥ 0. Then, we have
Proof 1. Since γ1, δ1 > 0, there exists a constant C > 0 such that τ(k) ≤ Cω(k) for . Using this fact, the definition of Gσ,s-norm and the estimate in Lemma 1, we obtain
Lemma 3. Let ψ(D) be the operator given in (6), σ > 0, and s > 1/2. Then, we have
Proof 2. Note that
Lemma 4. Let ψ(D) be the operator given in (6), σ > 0, and s ≥ 1. Then, we have
Proof 3. Using Lemma 1 with s − 1 ≥ 0, we have
Combining the inequalities found in Lemmas 2–4, we obtain the following nonlinear estimate.
Lemma 5. Let F(η) be given as in (8). Then, for any s ≥ 1 and σ > 0, we have
Now, applying the contraction mapping argument to the integral equation (9) and using Lemma 5 yields the following local result.
Theorem 2. Let η0 ∈ Hσ,s for s ≥ 1 and σ > 0. Then, there exists a unique solution
3. Approximate Conservation Law
Lemma 6. For all v ∈ H2, we have
To prove the estimates (44)–(46), we recall the following estimate from ([10], Lemma 3).
Lemma 7. Let p > 1 be any integer and such that |ξ1| ≤ |ξ2| ≤ ⋯≤|ξp| and . Let . Then,
Theorem 3. (Approximate conservation law). Let η0 ∈ Hσ,2 for all σ > 0, and η be the local-in-time solution to the IVP (1) that is constructed in Theorem 2. Then,
Then, plugging (61) and (62) into (42) and combining it with (41) yields the desired estimate (60).
4. Proof of Theorem 1
We closely follow the argument presented in [24, 30]. First, we consider the case s = 2. Then, the general case, s ≥ 1, will essentially reduce to s = 2 as shown in the next subsection.
4.1. The Case: s = 2
4.2. The General Case: s ≥ 1
This completes the proof of Theorem 1.
Conflicts of Interest
The author declares no conflicts of interest.
Funding
The author received no financial support for this manuscript.
Open Research
Data Availability Statement
Data sharing is not applicable to this article as no new data were created or analyzed in this study.