Positive solutions of critical semilinear elliptic equations on non-contractible planar domains (Q1592436): Difference between revisions

The author deals with the Dirichlet problem \[ Du= f(u)\quad\text{in }\Omega,\quad u= 0\quad\text{on }\partial\Omega,\tag{1} \] for a smooth, bounded domain \(\Omega\subset \mathbb{R}^2\) and a smooth function \(f:\mathbb{R}\to \mathbb{R}\) with polynomial growth. The solvability of equation (1) depends in a subtle way on the nonlinearity \(f(u)\) and not only on its growth at infinity. For instance if \(f(u)= u\exp(4\pi u^2)\), equation (1) admits a positive solution if the diameter of \(\Omega\) is sufficiently small, while if \(f(u)= u\exp(4\pi u^2- u)\), equation (1) does not admit any positive solution on any ball \(B_R(0)\) of radius \(R\) sufficiently small. For this reason the author calls these kind of problems with critical exponential growth, relating the criticality with the nonlinear elliptic problems with critical growth in \(\Omega\subset \mathbb{R}^n\), \(n\geq 3\), where the critical growth is given by the limiting exponent \(2^* = 2n/n- 2\) of the Sobolev embedding of \(H^1_0(\Omega)\) into \(L^p(\Omega)\). In this paper the author presents some results on the existence of positive solutions of (1) on noncontractible, smooth, and bounded domains \(\Omega\) of \(\mathbb{R}^2\), similar to the results of Coron and Bahri-Coron for equations with critical growth in dimensions \(n\geq 3\). He proves that positive solutions of (1) always exists for a large class of nonlinearities on a noncontactible domain of \(\mathbb{R}^2\). The proof of this result is quite delicate. The approach of Coron, which relies on an accurate analysis of the Palais-Smale sequences in critical growth problems in higher dimension, cannot be carried over to two-dimensional critical problems, because in two space dimensions the behavior of Palais-Smale sequences is more complicated than in the higher-dimensional case. Also other techniques as approximations with subcritical problems does not seem to work, because it appears to be impossible to control energy levels where the concentration of solution may occur. In the paper the strategy adopted is to approximate (1) with a one-parameter family of quasilinear elliptic equations. A delicate and accurate analysis of the behavior of a family of solutions of saddle type of the approximating problems, allows the author to prove the existence of a positive solution of (1) for some noncontractible domains.