Almost all linear spaces and partial \(t\)-designs have trivial automorphism groups (Q1336439): Difference between revisions

A linear space is denoted by \((S,\beta)\) where \(S\) denotes a set (elements of \(S\) called points) and \(\beta\) is a set of subsets of \(S\) (elements of \(\beta\) called lines) with the property that each pair of points belongs to a unique line. A \(t\)-design with blocks of size \(k\) is a pair \((S,\beta)\) where \(\beta\) denotes a set of \(k\) element subsets of \(S\) (subsets being called blocks) with the property that each subset of \(S\) with \(t\) elements is contained in a unique block. Such designs are denoted by \(s(t,k,n)\) and are also called Steiner systems. In case each subset with \(t\) elements is contained in at most one block, we call it then a partial Steiner system, and the \(t\)-subset not in any block is said to be uncovered. This paper is concerned with the automorphism groups of such structures. It is shown that each finite linear space has a trivial automorphism group (i.e. the identity permutation), and the same result holds for every partial \(s(t,t+1,n)\) when \(t\geq 2\). Finally, it is demonstrated that almost every \(s(t,t+1,n)\), \(t\geq 2\), has a trivial automorphism if, for \(\delta> 0\) and \(n\) very large, there are at least \(n^{{n\choose t}\left[{1\over t+1}-\delta\right]}\) Steiner systems \(s(t,t+1,n)\).