Let $A\to B$ be a full embedding of exact categories that induces an embedding $D^b(A)\to D^b(B)$. My question is: what can one say about the relation of the homotopy cofibre $K(A)\to K(B)$ (the relative K-theory) with the Verdier quotient $D=D^b(B)/D^b(A)$? Which additional stucture for $D$ could allow to define a certain K-theory for this triangulated category such that $K(D)\cong \operatorname{Cone} (K(A)\to K(B))$?
Could one recover the relative K-theory from the quotient derived category?
Mikhail Bondarko
- 16.9k
- 4
- 34
- 97