Completeness for concrete near-rings. (Q1887576)

The main result of this paper is to determine the maximal subnear-rings contained in \(M(\Gamma)\), the near-ring of all maps from a finite group \(\Gamma\) to itself. This is accomplished by using two distinct methods. If the maximal subnear-ring \(N\) contains the identity map then it is in one of three families: \(M_H:=\{f\in\Gamma^\Gamma\mid\forall x\in H\), \(f(x)\in H\}\) for some subgroup \(H\neq\Gamma\) of \(\Gamma\), \(M_{\Gamma:K}:=\{f\in\Gamma^\Gamma\mid\forall x,y\in\Gamma:x-y\in K\Rightarrow f(x)-f(y)\in K\}\) for some proper normal subgroup \(\{0\}\neq K\neq\Gamma\) of \(\Gamma\), \(M_{\text{aff}}:=\{f\in\Gamma^\Gamma\mid\forall x,y\in\Gamma:f(x+y)=f(x)+f(y)-f(0)\}\), where \(\Gamma\) is an elementary Abelian \(p\)-group. The method used here is to establish a relationship between near-rings and clones containing the group operations of the underlying groups which shows that the unary parts of such clones correspond precisely to near-rings containing the identity function. This enables the use of Rosenberg's characterization of maximal clones to obtain the result. When the maximal subnear-ring does not contain the identity function then it is of the form \(M_{a,b}:=\{f\in\Gamma^\Gamma\mid f(a)=f(b)\}\) for some \(a\neq b\) in \(\Gamma\). This result uses typical near-ring techniques. The characterization is then applied to determine conditions under which a subnear-ring of \(M(\Gamma)\) is the whole of \(M(\Gamma)\), which is of particular interest in determining single generators for \(M(\Gamma)\). As a consequence it is shown that the probability that a randomly chosen bijection from \(\Gamma\) to \(\Gamma\) generates \(M(\Gamma)\) is greater than \((n-6)/n\) for \(n\geq 6\), where \(|\Gamma|=n\). -- This is an interesting and thought-provoking paper.