Morse homology for the Yang-Mills gradient flow (Q441893): Difference between revisions

In this interesting paper, the author works out the analytic aspects of a Morse theory based on the Yang-Mills functional on bundles over Riemannian surfaces. The Yang-Mills heat flow can be considered as a flow on the connections modulo based gauge transformations, and this is a point of view which is suitable here. Given the set \({\mathcal P}(a)\) of based gauge equivalence classes of Yang-Mills connections of energy \(\leq a\) on a given bundle, a Morse function \(h:{\mathcal P}(a)\to{\mathbb R}\) is chosen. With these data, the author defines a Yang-Mills Morse complex, generated by the critical points of \(h\) and graded by a natural index. Now the main point of the paper is how to define a boundary operator for this complex, in order to define Morse homology. This boundary operator arises from the \(L^2\)-gradient flow of the Yang-Mills potential, modified by some generic perturbation term. The definition involves counting ``cascade configurations'' of gradient flow lines. The perturbation is necessary because it is not known whether the unperturbed Yang-Mills flow automatically obeys a Morse-Smale condition. Transversality is restored by choosing a suitable perturbation, the existence of which is everything but trivial. The main theorem of the paper states that all these data can actually be used to define the desired Morse homology, and that it is independent of the choice of perturbation and Morse function.