Timeline for Local index formula for >ungraded< elliptic operators
Current License: CC BY-SA 3.0
6 events
when toggle format | what | by | license | comment | |
---|---|---|---|---|---|
Dec 8, 2014 at 8:54 | vote | accept | AlexE | ||
Dec 8, 2014 at 8:54 | answer | added | AlexE | timeline score: 0 | |
Dec 3, 2014 at 8:27 | comment | added | AlexE | Yes, I assume that $P$ is essentially self-adjoint so that it indeed defines an element of $K_1(M)$. I also assume the manifold to be compact (and without boundary). Applying the homological Chern character (constructed by using Connes' cyclic cohomology), we get an element of $H_\ast(M)$ and I want to identify the Poincare dual of it. If $P$ is graded the Poincare dual is, up to some universal constants, $ind(P)$ which is computed / constructed using the clutching construction. This is a generalization of Atiyah-Singer. But what is this Poincare dual now in the ungraded case? | |
Dec 2, 2014 at 21:45 | comment | added | Johannes Ebert | Typically, you would replace the operator $P$ by $P \oplus P^{\ast}: E \oplus F \to E \oplus F$. This is graded, and the index (in the graded sense) is the usual index. An ungraded operator has an index in $K^1$, but only if it is self-adjoint. | |
Dec 2, 2014 at 16:25 | comment | added | Paul Siegel | There are a number of index theorems for operators whose K-homology classes live in the odd degree groups for whatever reason, but typically you need the operator to have some extra structure. Do you have a specific operator in mind? | |
Dec 2, 2014 at 16:17 | history | asked | AlexE | CC BY-SA 3.0 |