Morse functions for which the stationary phase approximation is exact (Q1092476)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Morse functions for which the stationary phase approximation is exact |
scientific article |
Statements
Morse functions for which the stationary phase approximation is exact (English)
0 references
1987
0 references
Let f be a Morse function with isolated critical points and distinct critical values on a compact manifold X. The well-known stationary phase method gives an approximation to the integral \(\int_{X}(\exp it f)dm\), \(t\to \infty\), by the sum \(\sum_{df(x)=0}(2\pi /t)^{n/2}(\exp it f(x))[\exp (n-2\lambda (x)\pi i/4]A_{x,k}(t)\). Here \(n=\dim X\), \(\lambda\) is index of f, for critical point x \[ A_{x,k}=\sum^{k}_{j=0}a_{x,j}(it)^{-j} \] \(a_{x,j}\) are real constants. The approximation is exact if the integral is equal to the sum with well-defined \(A_{x,\infty}(t)\). The main result of the paper under review is the theorem: If the stationary phase approximation for f is exact then the index of every critical point is even and its Morse inequalities are in fact equalities. This theorem has important consequences. If approximation for integral is exact then the homology of X is torsion-free and occurs only in even dimensions. In particular dim X is even [see also \textit{E. Witten}, J. Differ. Geom. 17, 661-692 (1982; Zbl 0499.53056)].
0 references
stationary phase approximation of integrals
0 references
Morse function
0 references