Construction of blow-up sequences for the prescribed scalar curvature equation on \(S^n\). II: Annular domains (Q1937831)

Throughout this work \(n\geq 3\), \(\widetilde{c}_n=\frac{n-2}{4(n-1)}\), \( \triangle_g\) is the Laplace-Beltrami operator associated with the metric \(g\) on \(S^n\), \( \triangle_0\) is the Laplace-Beltrami operator associated with the Euclidean metric \(g_0\) on \(\mathbb R^n\), \(\triangle_1\) is the Laplace-Beltrami operator associated with the standard metric \(g_1\) on \(S^n\), and \(R_g\) is the scalar curvature of \((S^n,g)\). The author considers the PDE \[ (1)\hskip 10pt \triangle_1u-\widetilde{c}_n\, n(n-1)\, u+(\widetilde{c}_n\, \mathcal{K})\, u^{\frac{n+2}{n-2}}=0 \] on \(S^n\). This is a generalization of the well-known Yamabe equation \[ (2)\hskip 10pt \triangle_gu-\widetilde{c}_n\, R_g\, u+\widetilde{c}_n\, n(n-1)\, u^{\frac{n+2}{n-2}}=0. \] The main objective is to find non-constant functions \(\mathcal{K}\) so that the equation (1) has an infinite number of positive solutions which compose a blow-up sequence of solutions. He applies the Lyapunov-Schmidt reduction method following a previous similar approach by Ambrosetti, Betti and Malchiodi to construct a \(C^k\)-metric \(g\) on \(S^n\) such that the Yamabe equation (2) has an infinite number of positive solutions. This is the second part of a work devoted to such a problem. Both parts are also supported by an appendix available at \url{http://www.math.nus.edu.sg/~matlmc/e-Appendix.pdf} (A similar version of this work can be found also in the \url{arXiv}). In the following we report the corresponding contents. Part I: 1. Introduction. 2. The Lyapunov-Schmidt reduction. 3. Tangent space, hyperbolic structure and the normal space. 4. Uniform cancelation. 5. Separation and blow-up. 6. First order derivation and the Kazdan-Warner condition. 7. Second derivation, symmetry and stability conditions. Part II: 1. Introduction. 2. Annular domains. 3. Infinite number of solutions. e-Appendix: A.1 A proof of integration by parts formula (2.11) in Part I. A.2 Riesz representation theorem. A.3 Non-degenerate critical points and fundamental notions in the degree theory. A.4 Curvature, gradient, Hessian, cancelation \& matching. A.5 Uniform approximation. A.6 Bounded projection. A.7 More uniform bounds. A.8 Argument toward Proposition 4.15 in Part I. A.9 \(G'\). A.10 Kazdan-Warner condition. A.11 Deriving (7.14) and (7.15), Part I. A.12 \(\lambda_M=\sqrt{(t+\triangle)(t-\triangle)}\) is the only critical point in expression (2.12), Part II. A.13 Deriving (2.39) in Part II. A.14 Back to \(S^n\) (refer to \(\S 3j\) in Part II). The author's goal is to show how, using annular domains, one can construct scalar curvature functions on \(S^n\) (\(n\geq 6\)), such that each of them enables the conformal scalar curvature equation to give a blowing-up sequence of positive solutions. The prescribed scalar curvature function has \(C^{n-1,\beta}\) regularity. Remark. It is useful to relate the problem considered in this work, studied with the tools of functional analysis, with the methods of the geometric theory of PDEs. In fact, one can apply to the equation (1) the same geometric methods utilized to study the Yamabe equation and the Yamabe problems in some previous works by the reviewer with \textit{R. P. Agarwal} [Adv. Math. Sci. Appl. 17, No. 1, 239--266 (2007; Zbl 1143.53017); Adv. Math. Sci. Appl. 17, No. 1, 267--285 (2007; Zbl 1140.53005)]. For this it is enough to assume that \(\mathcal{K}\) is a positive analytic function on \(S^n\). In these papers, a more general Riemannian manifold instead of \(S^n\) is considered. Then one can see that equation (1) is formally integrable and completely integrable. Therefore, in the neighborhood of any point \(q\) on this equation there exists an analytic solution. Furthermore, taking into account that the symbol is not zero, one can state the existence of singular solutions with singularity of Thom-Boardman type. These can be considered as solutions with blow-up points, in the sense of the present work. For Part I of the paper see [Commun. Contemp. Math. 14, No. 2, 1250008, 31 p. (2012; Zbl 1275.53039)].