Homoclinic and heteroclinic orbits for a semilinear parabolic equation (Q765667)

The authors ``study the existence of connecting orbits for the Fujita equation with a critical or supercritical exponent. For certain ranges of the exponent we prove the existence of heteroclinic connections from positive steady states to zero and a homoclinic orbit with respect to zero.'' Notably, the existence of homoclinic orbits may at first be surprising due to the presence of a free energy, but is readily explained as not lying in the energy space -- here the domain is the entire \(\mathbb{ R}^N\). Nevertheless, this seems to be the the first proof of a homoclinic solution for a (scalar) parabolic equation. The paper starts by a nice and concise discussion of the various critical exponents \(p\). Indeed, the details of the existence results in this paper require the exponent to lie in certain ranges. Specifically, the homoclinic solutions exist for exponents in the range between the Sobolev and Lepin exponents, which also means that \(N>2\). In fact, for \(N>10\) there even exist non-radial homoclinic solutions. The results also provide some sup-norm estimates, for instance, that of the homoclinic orbits behaves like \(t^{-1/(p-1)}\). In a nutshell, the method of proof relies on studying the linearization and singular homo- and heteroclinic solutions (which blow up but continue as weak solutions), and the comparison principle.