scientific article; zbMATH DE number 1098767

Publication date 22 April 1998

A big paper with a big surprise: nilpotency and solubility for groups and rings can be efficiently unified giving general penetrating results on \(\Omega\)-groups. Main tool is the near-ring \(N=P_0(V)\) of zero-fixing polynomial maps of an \(\Omega\)-group \(V\) into \(V\): so the ideals of \(V\) are its \(P_0(V)\)-subgroups and chain conditions have a good play. To make the paper self-contained the first seven sections collect (new and not so new) basic material, mainly linked to (\(n\))-tame near-ring theory. To exemplify the type of settings: an \(N\)-subgroup of the \(N\)-group \(V\) is called central in \(V\) if \((v+u)\alpha=v\alpha+u\alpha\), \(\forall v\in V\), \(u\in U\), \(\alpha\in N\). The centre \(Z(V)\) is the sum of all its central submodules and \(V\) is called \(N\)-nilpotent if there exists a finite chain of submodules \(\{V_i\}\) such that \(V_i/V_{i-1}\subseteq Z(V_{i-1})\) and similar definitions are given. If \(V\) is \(3\)-tame a Fitting submodule \(F(V)\) is defined using the analogously defined centralizers \(C_V(U/W)\) of minimal factors \(U/W\) of \(V\). Section nine starts to develope the main idea: commutators in groups and products in rings are both binary zero-symmetric operations. We will obtain such an operation starting by an \(\Omega\)-group \(V\) and faithfully representing the operations of \(\Omega\). So define a product of non-empty subsets \(S_1\), \(S_2\) of \(V\) as \[ (S_1,S_2)=\{(a_1,a_2)\varphi\mid (x,y)\varphi\in V_0[x,y],\;a_i\in S_i,\;i=1,2\}, \] where \(V_0[x,y]\) is the algebra of polynomials over \(V\) such that \((0,a)\varphi=(a,0)\varphi=0\), \(\forall a\in V\). Now define an annihilator \(A_V(B)\) for subsets \(B\) of \(V\) according to the new set product and have (e.g) the theorem: if \(B\) is a minimal ideal of \(V\) then \(A_V(B)=C_V(B)\) iff \(|B|\neq 2\) or \((B,B)=\{0\}\). Section 11 is a digression: If \(V\) is an associative ring with \(P_0(V)\) having DCCR then \(P_0(V)\) is finite. If moreover \(V\) has an identity, then \(V\) is finite. A study of annihilators leads to a multiplicative analogue \(F_m(V)\) of the Fitting submodule: it is an ideal and in the general case \(F_m(V)\subseteq F(V)\) but (calling abelian in \(V\) an ideal such that \((B,B)=\{0\}\)) if \(V\) has DCCI then either \(F_m(V)=F(V)\) or \(F(V)/F_m(V)\) is a finite direct sum of non-abelian minimal ideals of \(V/F_m(V)\) of order two. Now we are in the core of the paper: it is possible to fit usual group-theoretical solvability and nilpotence machinery using the new set product \((A,B)\) instead of ordinary commutator \([A,B]\), and so \(m\)-solvability and \(m\)-nilpotence are studied and linked (e.g) to \(F_m(V)\) and various chain conditions. Sample result: an ideal \(B\) of \(V\) is \(m\)-nilpotent in \(V\) iff it is \(m\)-solvable and \(N\)-nilpotent. To have pictures of relationship between unique maximal locally \(N\) (\(m\))- nilpotent (solvable) submodules \(L(V)\), \(L_m(V)\), \(S(V)\), \(S_m(V)\) of the \(\Omega\)-group \(V\) an ideal \(B\) of \(V\) is called \(2\)-constrained in \(V\) if there does not exist any non-abelian minimal factor \(C/D\) of \(V\), with \(C\leq B\) and \(|C/D|=2\). An example: \(S_m(V)\) is \(2\)-constrained in \(V\). Main theorem of section 18: \(L_m(V)\) is the intersection of \(L(V)\) and of \(C_2(V)\), the sum of all ideals of \(V\) which are \(2\)-constrained in \(V\). To end: let \(\text{Id}(A,B)\) be the ideal generated by \((A,B)\). The \(\Omega\)-group \(V\) is called \(m\)-perfect if \(\text{Id}(V,V)=V\), full if \(\text{Id}(V,B)=B\) for all ideals \(B\), \(e\)-rigid if all its \(N\)-endomorphisms are idempotent, rigid if for any ideal \(B\) it is \(\text{Id}(B,B)=B\). Sample results: A full \(\Omega\)-group \(V\) is \(e\)-rigid; if it has DCCI then the radical \(J(N)\) (i.e. \(J_2(N)\)) is nilpotent; if moreover \(V\) is \(N\)-nilpotent it is a finite discrete sum of non-abelian minimal ideals of order two. If \(V\) is rigid the lattice of its ideals is distributive; if moreover \(N/J(N)\) has DCCR then \(V\) and \(N\) are finite. Rigidity can be an alternative approach to semisimplicity. For the entire collection see [Zbl 0874.00036].

Recommendations

Cited in

(14)

This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4368941)