On translation transversal designs with \(\lambda >1\) (Q1080426)

An (s,k,\(\lambda)\)-translation divisible design is defined to be an incidence structure \({\mathcal I}=({\mathcal P},{\mathcal B})\) consisting of a finite set \({\mathcal P}\) of ''points'' and a set \(B\neq \emptyset\) of non-empty subsets of \({\mathcal P}\) of size k (called ''blocks'') with the following properties. (i) There is a partition of \({\mathcal P}\) into disjoint classes of size s (''point-classes'') such that any two points of different classes are joint by exactly \(\lambda\) blocks and such that there is no block joining different points of the same class. (ii) There exists an automorphism group G of \({\mathcal I}\) (called ''translation group'') acting regularly on \({\mathcal P}\) and satisfying \(b=b^ g\vee b\cap b^ g=\emptyset\) for every \(b\in {\mathcal B}\) and every \(g\in G.\) The case \(\lambda =1\) has been treated by several authors. The reviewed article investigates the situation for \(\lambda >1\). Using the classification of finite groups with a partition given by Baer, Kegel and Suzuki, there is shown the following theorem. Let \({\mathcal I}({\mathcal P},{\mathcal B})\) be an (s,k,\(\lambda)\)-translation divisible design with \(1<k<| {\mathcal P}|\) and \(\lambda >1\). Then the translation group G of \({\mathcal I}\) is either a p-group or a Frobenius group. For \(\lambda >1\), as well as for \(\lambda =1\), there are given examples of translation transversal designs (i.e. translation divisible designs having the property that each block meets each point class) which admit a p-group as a translation group. (The Frobenius group case will be treated in a separate paper.)