The well-posedness for the Camassa-Holm type equations in critical Besov spaces \(B_{p , 1}^{1 + \frac{ 1}{ p}}\) with \(1 \leq p < + \infty \) (Q6155347)

In this paper, the authors prove local well-posedness for the Camassa-Holm (CH) equation \[ u_t - u_{xxt} + 3uu_x = 2u_xu_{xx} + uu_{xxx}\tag{1} \] and the Novikov equation \[ u_t - u_{xxt} = 3uu_xu_{xx} + u^2u_{xxx} - 4u^2u_x\tag{2} \] in the critical Besov spaces \(\mathcal C([0,T]:B_{p,1}^{1+\frac 1p}(\mathbb R))\) with \(1 \leq p < \infty\). In addition, they prove local well-posedness for the two-component Camassa-Holm system \[ \begin{split} & u_t + uu_x = -\partial_x(1-\partial_{xx})^{-1}(u^2 + \frac 12 u_x^2 + \frac 12 \rho^2) \\ & \rho_t + u \rho_x = - u_x\rho \end{split}\tag{3} \] in \(\mathcal C([0,T]:B_{p,1}^{1+\frac 1p}) \times \mathcal C([0,T]:B_{p,1}^{\frac 1p})\). Their results follow by first establishing well-posedness results for the abstract equation \[ \begin{split} & u_t + A(u) u_x = F(u) \qquad t >0, x \in \mathbb R \\ & u(0,x) = u_0(x) \end{split}\tag{4} \] where \(A(u)\) is a polynomial and \(F\) is a 'good operator' such that for any \(\phi \in C_0^\infty\) and any \(\epsilon > 0\), the following fact holds. If \(u_n \phi \rightarrow u\phi\) in \(B_{p,1}^{1+\frac 1p-\epsilon}\), then \(\langle F(u_n),\phi \rangle \rightarrow \langle F(u),\phi \rangle\). In addition, they establish similar well-posedness results for an abstract system related to the two-component Camassa-Holm system. These results enable them to prove local well-posedness for (1), (2), and (3) in the Besov spaces discussed above. The local well-posedness results for the abstract equation (and abstract system) are obtained by using a Lagrangian coordinate transformation. Specifically, the associated Lagrangian scale of (4) is \[ \begin{split} & \frac d{dt}y(t,\xi)=A(u)(t,y(t,\xi)) \qquad t >0, \xi \in \mathbb R \\ & y(0,\xi)=\xi. \end{split} \] Then introducing the new variable \(U(t,\xi) = u(t,y(t,\xi))\), (4) becomes \[ \begin{split} & U_t = (F(u))(t,y(t,\xi)) := \widetilde{F}(U,y) \qquad t > 0, \xi \in \mathbb R \\ & U(0,\xi) = U_0(\xi) = u_0(\xi). \end{split}\tag{5} \] Establishing properties of (5) allow the authors to prove well-posedness results for (4) and thus (1) and (5). A similar technique is used to prove local well-posedness for the system (3).