On a periodic boundary value problem with exponential nonlinearities (Q1978225)

The main result here is the following theorem: The problem \[ -\ddot u =\lambda \left(2e^u/\int e^u -1\right),\quad u(-1)=u(1),\quad \dot u(-1)=\dot u(1),\quad \int u=0, \] has nontrivial solutions on the interval \(\langle -1, 1\rangle \) if and only if \(\lambda >\pi ^2\). Moreover, if \(\lambda >k^2\pi ^2\) holds for some \(k\in \mathbb{N}\) then this problem has at least \(k\) geometrically distinct solutions.