Table of Contents
Fetching ...

The zero-dispersion limit for the Benjamin--Ono equation on the circle

Ola Mæhlen

Abstract

Using the explicit formula of P. Gérard, we characterize the zero-dispersion limit for solutions of the Benjamin--Ono equation on the circle $\mathbb{T}= \mathbb{R}/2π\mathbb{Z}$ with bounded initial data $u_0\in L^\infty(\mathbb{T},\mathbb{R})$. The result generalizes the work of L. Gassot, who focused on periodic bell-shaped data, and complements the work of Gérard and X. Chen who identified the zero-dispersion limit on the line with $u_0\in L^2\cap L^\infty(\mathbb{R})$. Here, as well as in the mentioned cases, the characterization agrees with the one first obtained by Miller--Xu for bell-shaped data on the line: The zero-dispersion limit is given as an alternating sum of the branches of the multivalued solution of Burgers' equation. From this characterization, we compute regularity properties of the zero-dispersion limit, including maximum principles and an Oleinik estimate.

The zero-dispersion limit for the Benjamin--Ono equation on the circle

Abstract

Using the explicit formula of P. Gérard, we characterize the zero-dispersion limit for solutions of the Benjamin--Ono equation on the circle with bounded initial data . The result generalizes the work of L. Gassot, who focused on periodic bell-shaped data, and complements the work of Gérard and X. Chen who identified the zero-dispersion limit on the line with . Here, as well as in the mentioned cases, the characterization agrees with the one first obtained by Miller--Xu for bell-shaped data on the line: The zero-dispersion limit is given as an alternating sum of the branches of the multivalued solution of Burgers' equation. From this characterization, we compute regularity properties of the zero-dispersion limit, including maximum principles and an Oleinik estimate.

Paper Structure

This paper contains 12 sections, 12 theorems, 86 equations, 2 figures.

Key Result

Theorem 1.1

Let $u_0\in L^\infty(\mathbb{T},\mathbb{R})$ and $u^\varepsilon\in C(\mathbb{R}, L^2(\mathbb{T},\mathbb{R}))$ denote the unique solutionIn the sense of Molinet MolinetPeriodic or, equivalently, Gérard--Kappeler SharpWellPosednessCircle. of eq: BOepsilon. Then, as $\varepsilon\to 0$, $u^\varepsilon$ locally uniformly in $t\in\mathbb{R}$. The function $ZD[u_0](t)$ is, for a.e. $x\in \mathbb{T}$, gi

Figures (2)

  • Figure 1: The graph (black curve) of the multivalued solution $u_B=u_0(x-2t u_B)$ of Burgers' equation, shown at $t=0$ and at a later time $t>0$. The red region corresponds to $\chi(u_0(x-2ty),y)=1$, while the blue region corresponds to $\chi(u_0(x-2ty),y)=-1$. The dashed vertical line intersects the graph once initially and seven times at the later time; the alternating sum of these seven $y$-values coincides with the integral of $\chi(u_0(x-2ty),y)$ along the dashed line.
  • Figure 2: A visual representation of the twelve weights $S(T^j m,T^\ell n)$ resulting from $m=(0,1,1,1)$ and $n=(0,1,2)$. The diagram at column $j$ and row $\ell$ corresponds to $S(T^j m,T^\ell n)$. The red and blue line represents the partial sums of $T^j m$ and $T^\ell n$ respectively; the former grows vertically and the latter horizontally. Moreover, $S(T^j m,T^\ell n)=1$ if the red line is below the blue, and $S(T^j m,T^\ell n)=0$ otherwise, as justified by Lemma \ref{['lem: aSimplerCharacterizationOfWhenTheCompositionOfShiftsIsOne']}. Here, we see that $S(T^j m,T^\ell n)=1$ when $(j,\ell)\in\{2,3,4\}\times\{3\}$.

Theorems & Definitions (30)

  • Theorem 1.1: Characterizing the zero-dispersion limit
  • Theorem 1.2: Properties of $ZD{[\cdot]}$
  • Lemma 2.1: Compositions of Toeplitz-shifts
  • proof
  • Proposition 2.2: Existence of a zero-dispersion limit
  • proof
  • Remark 2.3: On the restriction $u_0\in L^\infty(\mathbb{T},\mathbb{R})$
  • Remark 2.4: Letting the initial data depend on $\varepsilon$
  • Proposition 2.5: Generalized alternating sum formula
  • Remark 2.6
  • ...and 20 more