Derived categories of toric varieties. III (Q253160): Difference between revisions

Following the author's previous works [Mich. Math. J. 54, No. 3, 517--535 (2006; Zbl 1159.14026); ibid. 62, No. 2, 253--363 (2006; Zbl 1322.14075)], this article addresses the derived McKay conjecture, stating an affirmative answer for finite abelian actions \(G\subset\)GL\((n,\mathbb{C})\) on \(\mathbb{C}^n\). More precisely, if \(f:Y\to X=\mathbb{C}^n/G\) is a \(\mathbb{Q}\)-factorial terminal relative minimal model of \(X\), then there exist toric closed subvarieties \(Z_i\subset X\) with \(Z_i\neq X\) for \(1\leq i\leq m\) and \(m\geq0\), such that there is a semi-orthogonal decomposition: \[ D^b(coh([\mathbb{C}^n\!/G])) \cong \mathrm{Span}{D^b(coh(\widetilde{Z}_1)),\ldots,D^b(coh(\widetilde{Z}_m)),D^b(coh(\widetilde{Y}))} \] where \(\widetilde{Y}\) and \(\widetilde{Z}_i\) are smooth Deligne-Mumford stacks associates to \(Y\) and \(Z_i\) respectively. Moreover, if \(G\subset\)SL\((n,\mathbb{C})\) then \(m=0\) thus recovering the ``Classical'' derived McKay correspondence conjecture in this case. For \(n=2\) the author is able to remove the abelian condition, so in this case it is proved that for any \(G\subset\)GL\((2,\mathbb{C})\) there is a semi-orthogonal decomposition: \[ D^b(coh([\mathbb{C}^2\!/G])) \cong \mathrm{Span}{D^b(coh(Z^\nu_1)),\ldots,D^b(coh(Z^\nu_m)),D^b(coh(Y))} \] where \(Z^\nu_i\) are normalizations of \(Z_i\). If in addition the group \(G\) is small, i.e.\ it has no quasireflections, then the semi-orthogonal complement of \(D^b(coh(Y)\) in \(D^b(coh([\mathbb{C}^2\!/G]))\) is generated by an exceptional collection. The paper concludes with the proof of the following result: Any proper birational morphism \(f:(X,B)\dashrightarrow(Y,C)\) between toric pairs which is a \(K\)-equivalence (i.e.\ \(K_X+B=K_Y+C\)), is decomposed into a sequence of flops. If in addition the coefficients of \(B\) belong to the estandar set \(\{1-1/m:m\in\mathbb Z_{>0}\}\) then \(D^b(coh(\widetilde{X})) \cong D^b(coh(\widetilde{Y}))\), thus proving in this case the ``\(K\) implies \(D\) conjecture'' proposed in [the author, J. Differ. Geom. 61, No. 1, 147--171 (2002; Zbl 1056.14021)].