A generalization of von Neumann regularity (Q2488273)

The author unifies results on certain (Rumely) domains, arising from the work of \textit{L. van den Dries} and \textit{A.~Macintyre} [``The logic of Rumely's local-global principle'', J.~Reine Angew.~Math. 407, 33--56 (1990; Zbl 0703.13021)] and \textit{A.~Prestel} and \textit{J.~Schmid} [``Existentially closed domains with radical relations'', J.~Reine Angew.~Math. 407, 178--201 (1990; Zbl 0691.12013)], on the theory of the ring of absolute algebraic integers, with earlier results of \textit{A.~Carson} [``The model-completion of the theory of commutative regular rings'', J.~Algebra 27, 136--146 (1973; Zbl 0279.02037)] and \textit{L.~Lipshitz} and \textit{D.~Saracino} [``The model companion of the theory of commutative rings without nilpotent elements'', Proc.~Am.~Math.~Soc. 38, 381--387 (1973; Zbl 0267.02040)] on the theory of commutative von Neumann regular rings, especially as approached \textit{via} sheaves of localisations over the maximal ideal spectrum [see also \textit{G.~Cherlin}, Model theoretic algebra. Selected topics. Lect. Notes Math. 521, Springer-Verlag, Berlin (1976; Zbl 0332.02056)]. For example, commutative rings \(R\) which satisfy the following properties are considered: every non-constant monic polynomial over \(R\) has a root in \(R\); every finitely generated ideal is principal; any finite system of polynomial equations which is solvable in the localisation of \(R\) at \(a\) and in the localisation of \(R\) at \(b\) is solvable in the localisation of \(R\) at \(d\) where \(aR+bR=dR\); for all non-zero \(a,b\in R\) there are \(a_1, a_2\in R\) such that \(a=a_1a_2\), \(a_1\) and \(b\) are relatively prime and \(b\in \text{rad}(a_2)\); there is a non-zero non-unit and every non-zero non-unit is the product of two relatively prime non-units; the set of maximal ideals \(M\) of \(R\) such that the localisation of \(R\) at \(M\) is a field is open in the Zariski topology on \(\text{ Maxspec}(R)\); there are no non-zero nilpotent elements; for each \(a\in R\) the set of maximal ideals \(M\) such that the image of \(a\) in the localisation of \(R\) at \(M\) equals \(0\) is Zariski-closed. This, and variants, are shown to be axiomatisable and to be model-complete. The proofs make heavy use of the representation of \(R\) and its localisations as rings of sections over the maximal ideal spectrum. The author has established a natural context for these results, with the original motivating examples lying at opposite extremes.