Compactness issues and bubbling phenomena for the prescribed Gaussian curvature equation on the torus (Q889717): Difference between revisions

Let \(M\) be a closed, connected Riemannian surface, whose Euler characteristic \(\chi(M)\) is zero and let \(g_0\) be a smooth background metric with identically vanishing Gauss curvature \(K_{g_0}\). If \(g=e^{2u}g_0\) is a conformal metric then \(K_g=-e^{-2u}\Delta_{g_0}u\). Let \(f_0:M\to\mathbb R\) be a smooth non-constant function with \(\max_{p\in M}f_0(p)=0\), all of whose maximum points are non-degenerate, and let \(f_\lambda:=f_0+\lambda\), where \(\lambda\in\mathbb R\). A natural question appears: for which values of \(\lambda\) the function \(f_\lambda\) is the Gauss curvature of a metric conformal to \(g_0\). That is equivalent to ask for which values of \(\lambda\) the equation \[ -\Delta_{g_0}u=f_\lambda e^{2u}\eqno{(1)} \] admits a solution. It is known that the above equation has a solution if and only if \[ \int_M f_\lambda d\mu_{g_0}=\int_M f_0 d\mu_{g_0}+\lambda<0 \] and \(f_\lambda\) is sign changing. Thus, the equation (1) is solvable if and only if \[ \lambda\in\Lambda:=(0,\lambda_{\max}),\quad \lambda_{\max}:=-\int_M f_0 d\mu_{g_0}. \] In the reviewed paper the author studies the behaviour of the set of solutions of (1) when \(\lambda\) tends to either \(0\) or \(\lambda_{\max}\).