Gradient flows of Morse-Bott functions (Q1592358): Difference between revisions

Recently \textit{F. R. Harvey} and \textit{J. B. Lawson jun.} presented a new unifying approach to the Morse theory based on the consistent use of the geometric measure theory [Ann. Math. (2) 153, No.~1, 1-25 (2001; Zbl 1001.58005)]. They consider the following question. For a given smooth flow \(\phi_t:X\to X\) on a manifold \(X\), when does the limit \(P(\alpha)=lim_{t\to\infty}\phi_t^*(\alpha)\) exist for a given smooth differential form \(\alpha\)? For a so called \textit{finite-volume flow} they show the existence of the limit for all \(\alpha\) and continuity of the corresponding linear operator \(P:{\mathcal E}^*(X)\to {\mathcal D'}^*(X)\) from smooth forms to generalized forms, i.e., currents. The operator \(P\) is chain homotopic to the inclusion \(I:{\mathcal E}^*(X)\to {\mathcal D'}^*(X)\) where the homotopy \(I-P=d\circ T + T\circ d\) is given with the help of a continuous operator \(T:{\mathcal E}^*(X)\to {\mathcal D'}^*(X)\) of degree \(-1\) constructed with the help of the graph of the flow. In the paper by Harvey and Lawson the finite-volume flow is the gradient flow of a Morse function \(f:X\to R\). In the paper under review the author generalizes this paper to the context of Morse-Bott functions \(f:X\to R\). He proves the existence, obtains explicit formulas for \(P\) and a geometric model for a Morse complex associated to \(f\) which might be used to compute the homology of \(X\) with \(\mathbb{Z}\) or \(\mathbb{Z}_2\) coefficients.