4
$\begingroup$

Any one knows a reference for computing K_0 of Algebra of zeroth order Pseudo's on a closed manifold in terms of explicit generators?

Thanx!

$\endgroup$

1 Answer 1

4
$\begingroup$

I do not have a reference, except that Higson-Roe's book ''Analytic K-homology'', p. 47ff should give enough information.

Let $P(M)$ be the algebra of order zero scalar pseudodifferential operators on the closed manifold $M$. There is a short exact sequence

$$ K(L^2(M)) \to P(M) \to C(SM) $$

where $C(SM)$ are the functions on the unit sphere tangent bundle. The long exact sequence in $K$-theory is

$$0=K_1 (K(L^2(M))) \to K_1 (P(M)) \to K_1 (C(SM)) \to K_0 (K(L^2 (M)))=\mathbb{Z} \to K_0 (P(M)) \to K_0 (C(SM)) \to 0.$$

I claim that the map $K_1 (C(SM)) \to \mathbb{Z}$ is surjective. The map sends a unitary matrix $s$ with entries in $C(SM)$ to the index of a pseudodifferential operator with symbol $s$, so the surjectivity amounts to saying that on each clsed manifold, there is a pseudo-DO with index $1$ and unitary symbol. Such an operator is constructed in Atiyah-Singer ''Index of elliptic operators I'', using the excision principle for the index.

The answer is that $K_0 (P(M)) \to K_0 (C(SM))$ is an isorphism; the group $K_0 (C(SM))$ of course depends heavily on $M$.

$\endgroup$
1
  • $\begingroup$ It seems to me that this algebra P(M) defined in Higson-Roe's book contains only the classical PDOs (i.e., with homogeneous symbol)? $\endgroup$
    – AlexE
    Commented Jan 7, 2014 at 9:22

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .