Kawahara equation in a quarter-plane and in a finite domain ∗ †

This work is concerned with the existence and uniqueness of globalin-time regular solutions for the Kawahara equation posed on a quarter-plane and on a finite domain.


Introduction
This work is concerned with the existence and uniqueness of global-in-time regular solutions for the Kawahara equation posed on a quarter-plane and on a finite domain.Our study is motivated by physics and numerics: the nonlinear relation, called Kawahara equation [19], is the fifth-order dispersive-type partial differential equation describing one-dimensional propagation of small amplitude long waves in various problems of fluid dynamics and plasma physics [2,30].This equation is also known as perturbed KdV or the special version of the Benney-Lin equation [3,4].
Methods to study initial and initial-boundary value problems for the KdV and Kawahara equations are similar, but differ in details for three types of problems, namely: the pure initial value problem (see [4,20,13,28] and the references); initialboundary value problems posed on a finite interval (see [8,9,10,14,15,17,22,21,25,26,27]); and problems posed on quarter-planes which is the case that attracts here our interest.For the KdV equation there is rather developed theory of well-posedness for such problems, see, for instance, [5,6,7,11,18,29].On the other hand, we do not know any published results dealing with well-posedness of the initial-boundary value problems for the Kawahara equation in quarter-planes and in finite domains.
The main results of this paper are the existence and uniqueness of a global regular solution to the initial-boundary value problems posed on the quarter-plane {x ∈ R + , t ≥ 0} and on the finite domain {x ∈ (0, 1), t ≥ 0} for the nonlinear Kawahara equation.Moreover, we show that the rate of decay of the obtained solution in the quarter-plane as x → ∞ does not depend on t > 0 and is the same as one of the initial data.Moreover, we prove that solutions to an initial-boundary value problem for the KdV equation can be approximated by corresponding solutions to initial-boundary value problems for the Kawahara equation.To prove these results we propose the transparent and constructive method of semi-discretization with respect to t which can be used for numerical simulations.
Furthermore, to obtain necessary a priory bounds in a quarter-plane, in place of usual for the KdV-type equations "artificial" weights such as 1 + x or e x/2 (see citations above), we use the "natural" exponential weight e kx where k > 0 is the decay rate of the initial data.This brings technical difficulties, but compensates in obtaining the same decay rate of the solution, while x → ∞, as one of the initial data which seems to be a new qualitative property.
It should be noted also that imposed boundary conditions are reasonable both from physical and mathematical point of view, see [7,30] and comments in [15].
To prove these results, first we solve a corresponding stationary problem exploiting the method of continuation with respect to a parameter.Then we prove solvability of a linear evolution problem by the method of semi-discretization with respect to t.Using the contraction mapping arguments, we obtain a local in time regular solution to the nonlinear problem.Finally, necessary a priori estimates allow us to extend the local solution to the whole interval t ∈ (0, T ) with arbitrary T > 0.

For real
In Q T we consider the one-dimensional nonlinear Kawahara equation (see [19]) subject to initial and boundary conditions Here and henceforth u : R + × [0, T ] → R is a unknown function, u t denotes its partial derivative with respect to t > 0, D j = ∂ j /∂x j are the derivatives with respect to x ∈ R + of order j ∈ N and u 0 (x) Throughout the paper we adopt the usual notations • and (•, •) to denote the norm and the inner product in L 2 (R + ).Symbols C, C 0 and C i , i ∈ N, mean positive constants appearing during the text.
The main results of this article are the following.

Stationary problem
Our purpose in this section is to solve the stationary boundary value problem Here a > 0 is a constant coefficient, u : R + → R is an unknown bounded function, D m denotes, as earlier, the m-th derivative with respect to x and f (x Theorem 2 Let a > 0 and f satisfies (3.3).Then (3.1),(3.2) admits a unique solution u ∈ H 5 (R + ) such that We start from the simpler equation subject to boundary data (3.2).Its general solution can be easily found by standard methods of ODE.

, N, and
We approximate (4.1)-( 4.3) by the following system: Due to results on solvability of the boundary value problem for the stationary equation, we can prove necessary a priori estimates for u n which imply the following Theorem 3 Let u 0 (x) and f (x, t) satisfy (4.4) and (4.5).Then (4.1)-( 4.3) admits a unique solution 5. Nonlinear problem.Local solutions.
Using the contraction mapping principle, we prove the existence and uniqueness of local regular solutions to the following nonlinear problem: (5.1) Then there exists a real T > 0 such that (5.

Global solutions
A priori estimates uniform in t ∈ (0, T ) now allow us to extend the obtained local solution to the whole (0, T ) with arbitrary fixed T > 0, hence to prove Theorem 1. Remark.Making use of the approach exploited to prove Theorem 1 and results from [15], we can solve the Kawahara equation posed on a finite interval: u(x, 0) = u 0 (x), x ∈ (0, 1).(6. 3) The following assertions are valid.