Sheaves on toric varieties for physics (Q1294799)

Nowadays theoretical physicists are more and more interested in all kinds of moduli spaces. The underlying paper gives a description of so-called equivariant sheaves on toric varieties. Much of the theory is due to A.~Klyashko. One may apply the technology to heterotic compactifications, in particular to the \((0,2)\)-models of Distler and Kachru. Let \(X\) be a toric variety with action of the algebraic torus \(T\). The action \(t:X\rightarrow X\), \(t\in T\), induces one \(\mathcal E\rightarrow t^*\mathcal E\) for any sheaf \(\mathcal E\) on \(X\). By definition, the fixed point locus of this induced \(T\)-action on the moduli space of sheaves (or bundles) on \(X\) consists of the \textit{equivariant sheaves}. Following ideas of Klyashko, one describes equivariant bundles on a toric variety by their behavior near toric divisors. In this way one is led to specify bundles in terms of a family of filtrations of a vector space, one for each toric divisor. As an application one may describe Chern classes by means of these filtrations. The theory is extended to torsion free sheaves on arbitrary toric varieties. For singular varieties one must distinguish between Cartier and Weil divisors. The global \(\text{Ext}\) groups for equivariant torsion free sheaves on a smooth toric variety can be calculated. An illuminating example on \({\mathbb P}^2\) illustrates some of the ideas behind. To come to terms with the moduli spaces of equivariant sheaves one is naturally led to a notion of (semi-)stability, well known from GIT (Mumford's geometric invariant theory). The moduli spaces of equivariant reflexive sheaves, equivariant bundles are equivariant torsion free sheaves and their relations are briefly discussed. Also, the moduli space of `not-necessarily-equivariant' sheaves, Bialynicki-Birula stratifications, monads and bundles on projective spaces, and as a byproduct the famous ADHM construction for instantons, are described. The paper closes with appendices on algebraic groups, sheaf classifications, GIT quotients, and filtrations. It is remarkable that nowhere is explicitly mentioned a proposition or a theorem, and that relevant mathematical material is deferred to footnotes. The paper should do well as a chapter in a book on (sheaves on) toric varieties.