2
$\begingroup$

Let $E$ be a vector bundle over a compact connected Hausdorff space $X$. To an endomorphism $\alpha \in End(E)$, we associate a $C(X)-$ module $\Gamma(E,\alpha)$ consisting of all $\beta\in End(E)$ such that there exists a $\gamma \in End(E)$ with $\beta=\alpha \circ \gamma$.

$$\require{AMScd} \def\diaguparrow#1{\smash{\raise.6em\rlap{\scriptstyle #1} \lower.3em{\mathord{\diagup}}\raise.52em{\!\mathord{\nearrow}}}} \begin{CD} && E\\ & \diaguparrow{\exists \gamma} @VV \alpha V \\ E @>> \beta> E \end{CD}$$

Is $\Gamma(E,\alpha)$ necessarily a projective $C(X)-$ module? Is it necessarilly a finitly generated module? Is it true to say that $\Gamma(E, \alpha)$ is a finitely generated projective module if and only if $\alpha$ is a strict morphism, namely $\dim( \ker \alpha_x)$ is independent of $x\in X$?

Remark 1: If $\dim( \ker \alpha_x)$ is independent of $x\in X$, then $\Gamma(E,\alpha)$ is a finitely generated projective module. Because it is isomorphic to $hom(E,F)=\Gamma (Hom(E,F))$, the module of sections of the Home bundle $Home (E,F)$ where $F=\text{Range}\; \alpha$.

Remark 2: in this question we actually work with a pair $(E, \alpha)$ in the category of vector bundles and considered the module of "Liftable" morphisms $\beta$. One cane repeat the same consideration but on other categories. For examples, in the category of Banach space, one obtain a Banach algebra $\Gamma(E, \alpha)$ associated to a pair $(E, \alpha)$ where $\alpha$ is a bounded operator on Banach space $E$. So Is there a categorical terminology for such construction?

$\endgroup$
4
$\begingroup$

It seems that your $C(X)$-module $\Gamma(E,\alpha)$ is just the image of $\alpha_*\colon \hom(E,E)\rightarrow \hom(E,E)$. The module $\Gamma(E,\alpha)$ is a direct summand of $\Gamma(E',\alpha')$ for $\alpha'\colon E'\rightarrow E'$ an extension of $\alpha$ to a trivialization $E\subset E'$ of $E$. Therefore, we can assume without loss of generality that $E$ is a trivial bundle. Then $\Gamma(E,\alpha)$ is a direct sum of $n=\dim E$ copies of the $n\times n$ matrix in $C(X)$ induced by $\alpha$, which can be anyone. Therefore the cokernel of $\alpha_*$ is the direct sum $n$ copies of any given finitely presented $C(X)$-module. This proves that your first question is equivalent to whether any finitely presented $C(X)$-module has projective dimension $\leq 1$. This would mean that the weak global dimension of $C(X)$ is $\leq 1$. I don't know much about these things and haven't been able to find much in internet about this topic, but you may open here a question on the weak global dimension of function rings. The answer to the second question is positive, as we have just seen.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.