Verma type modules of level zero for affine Lie algebras (Q4337952)

In contrast to the case of simple finite-dimensional complex Lie algebras, not all partitions of the root system of an affine Lie algebra \(\mathfrak g\) into a disjoint union of closed subsets are equivalent. With any partition as above one associates a Borel subalgebra \(B\) of \(\mathfrak g\) and thus one can define an analogue of Verma module. There is one partition, which is called standard and which leads to the standard notion of Verma module over \(\mathfrak g\). Given a partition, which is not equivalent to the standard one, the corresponding Borel subalgebra is called non-standard and the corresponding analogue of the Verma module is called the Verma type module. The Verma type modules of level zero (i.e. such that the center of \(\mathfrak g\) acts trivially) are the main objects of interest in the paper under review. The author also defines a notion of a non-standard parabolic subalgebra and studies the corresponding analogue of generalized Verma modules. The main results of the paper are the following: 1. The author describes all simple subquotients of the Verma type module \(M\) of zero level, satisfying a special condition, that for the naturally defined submodule \(N\), the module \(M/N\) is simple. In the other case (\(M/N\) is not simple) an analogous result is established only for ``almost all'' Verma type modules. Additionally, the author reduces the problem of reducibility of \(M/N\) to the problem of reducibility for its finite part (the sum of all finite-dimensional weight spaces), which is, in fact, a Verma module over a subalgebra of \(\mathfrak g\). 2. The author constructs a strong BGG-resolution for the unique simple quotient of a Verma type module in several cases. 3. The author defines certain categories of \(\mathfrak g\)-modules in which Verma type modules and their simple quotients are natural objects and proves the BGG-reciprocity for these categories. 4. The author describes the structure of some generalized Verma type modules. \{Reviewer's comment: There are some misprints in the paper, for example in the proof of Proposition 5.2 the author refers to the same Proposition\}.