Concentration of solutions for the fractional Nirenberg problem (Q906810)

The author studies the presentation of a solution for the fractional problem \[ (*)_{K(\cdot)}:\;(-\Delta)^su(\cdot)=K(y)[u(\cdot)]^{\frac{n+2s}{n-2s}} \text{ in } \mathbb R^n \text{ such that } u(\cdot)>0, \] where \(s\in (0,1)\)\textbf{,} \(K(\cdot)\) is a twice continuously differentiable bounded function in \(\mathbb R^n\), and \((-\Delta)^s\) represents the typical fractional Laplacian operator. Accordingly, by imposing meaningful conditions on \(K(\cdot)\), the author states that a solution of \((*)_{K(\cdot)}\) takes the form \[ u(\cdot)=\sum_{j=1}^2K_{i_j}^{\frac{2s-n}{4s}}U_{x^j,\lambda_j}(\cdot)+v(\cdot), \] such that \(K_{i_j}\) are some constants and for \(j\in\{1,2\}\), the functions \(U_{x^j,\lambda_j}(\cdot)\) are (almost) explicit solutions for \((*)_0\) with \((x^j,\lambda_j)\in \mathbb R^n\times (\mu^{-1},\infty)\) (\(\mu\) is some small constant), the distance between \(x^1\) and \(x^2\) (resp. \(x^j\) and \(z^{i_j}\) (a local maximum of \(K(\cdot)\)) is greater than or equal to \(\mu^{-1}\) (resp. less than or equal to \(\mu\)), and \(v\) pertains to an explicit orthogonal subspace in \(H^s(\mathbb R^n)\), the fractional Sobolev space, as well the \(H^s(\mathbb R^n)\)-norm of \(v\) is bounded by \(\mu\) (Theorem 1.1). The proof of this latter theorem is mainly based on the application of the calculus of variations, that is, on looking for the critical points of the energy functional associated to \((*)_{K(\cdot)}\).