On structures of modular adjacency algebras of Johnson schemes (Q536222): Difference between revisions

The aim of this short paper is to consider the structure of the adjacency algebra of the Johnson scheme, primarily in prime characteristic. Recall that an association scheme consists of a set \(X\) and a partition of \(X \times X\) into relations, which satisfy certain homogeneity properties. The Johnson scheme \(J(m,n)\) takes \(X\) to be the set of \(n\)-subsets of a fixed \(m\)-set, with relations \(R_0,\dots,R_n\), where \((A,B) \in R_i\) if and only if \(|A\cap B|=n-i\). If \(F\) is a field, then the adjacency algebra of an association scheme is simply the algebra generated by the adjacency matrices of the relations in the scheme. The paper under review presents several results about the structure of the adjacency algebra \(FJ(m,n)\), when \(F\) has characteristic \(p\). These are mostly concerned with the decomposition of this algebra into blocks, and the structures of these blocks in certain cases. However, the paper is not at all well written; it looks as though it was written in great haste. The author crams his results into six pages, but would have been much better advised to spend some time setting the scene, and recalling background details. As it is, the reader is expected to know the definitions (and motivation) concerning association schemes, and to have some understanding of the structure of commutative algebras. The results which are cited from elsewhere (which comprise most of the paper) are spread through the paper and the whole is very poorly organised. There is no indication given of which are the main, or most important, results. Furthermore, the author's use of mathematical terminology is sloppy at times (for example, he talks about an algebra being `a finite-representation type'). In summary: the paper may prove to be a helpful reference for someone who needs the structure of an \(FJ(m,n)\). But it should not be read by someone trying to learn about the subject in general.