Complete classification and nondegeneracy of minimizers for the fractional Hardy-Sobolev inequality, and applications (Q2656282)

In this paper the authors studied firstly the problem \[\begin{cases} (-\Delta)^su=|x|^{-bq}u^{q-1} & u\in \mathcal{D}^s(\mathbb{R}^n) \\ u>0.\tag{P0}\end{cases}\] where \(n\geq 2\), \((-\Delta)^s\) is the fractional Laplacian of order \(s \in (0, 1)\) and \(q,b\) satisfy \[2<q<2_s^*=\frac{2n}{2-2s}, \frac{n}{q}-b=\frac{n}{2}-s.\] Let \(z_1\in \mathcal{D}^s(\mathbb{R}^n)\) be the positive and semicontinuous minimizer for the best constant \(S_q\) satisfying \[S_q\||x|^{-b}u\|_q^2\leq \|(-\Delta)^{\frac{s}{2}}\|_2^2,\ \forall u\in \mathcal{D}^s(\mathbb{R}^n),\] and let \(z_t(x)=t^{\frac{2s-n}{2}}z_1(\frac{x}{t})\) for \(t>0\). The first theorem claims that: (i) if \(u\in \mathcal{D}^s(\mathbb{R}^n)\) is a solution to \((\mathcal{P}0)\) then \(u = z_t\) for some \(t > 0\); (ii) \(z_1\in \mathcal{C}^\infty(\mathbb{R}^n\backslash\{0\})\cap L^\infty(\mathbb{R}^n)\cap \mathcal{C}^\alpha(\mathbb{R}^n)\) for any \(\alpha\in [0, 2s-bq)\) and there exist positive constants \(C_1, C_2\) such that \(\frac{C_1}{1+|x|^{n-2s}}\leq z_1(x)\leq \frac{C_2}{1+|x|^{n-2s}}\). The second theorem states that any solution \(v\in \mathcal{D}^s(\mathbb{R}^n)\) to the linearized problem: \((-\Delta)^sv=(q-1)|x|^{-bq}z_t^{q-2}v\) is proportional to \(\dot{z}_t\). As a consequence of the second theorem, they obtained a symmetry result for ground state (i.e. least energy) solutions to the nonlocal problem \((-\Delta)^su+\lambda|x|^{-2s}u=|x|^{-bq}u^{q-1},\ u\in \mathcal{D}^s(\mathbb{R}^n)\). As a further consequence of the second theorem, they obtained by a Lyapunov-Schmidt argument sufficient conditions on a prescribed weight \(k(x)\) which guarantee the existence of solutions to the perturbative problem: \((-\Delta)^su=(1+\epsilon k(x))|x|^{-bq}u^{q-1},\ u\in \mathcal{D}^s(\mathbb{R}^n), \ u>0\).