Asymptotic stability at infinity for bidimensional Hurwitz vector fields (Q2435186)

The paper deals with flows on the plane which was compactified by adding the point at infinity and making it an equilibrium point. When the flow is linear and all eigenvalues have negative real parts, it is evident that infinity is a repeller. For a non-linear field, it is called Hurwitz provided that its linearization at every point has all eigenvalues negative. At first sight it seems to be impossible to say anything globally based on such a hypothesis. Nevertheless, the paper demonstrates, among other things, that the property of infinity being an attractor or repeller can be obtained under rather general conditions. They include the absence of fixed points other than infinity, which for topological reasons also requires excluding a neighborhood of 0 from the domain of definition. The most striking result is that then one can add a constant to the field, and then infinity becomes either an attractor or a repeller. Not surprisingly the proof has a flavor of differential topology, more so than of dynamical systems. The reviewer is not a specialist in this kind of problems and believes that few people are. I have to add that while being impressed with this clearly deep result, I was also disappointed to find the paper rather hermetic. For example, the intricate topological definition of an attractor given in the paper eluded me and forced me to rely on an intuitive grasp.