On a problem in poe-semigroups (Q809108)

A poe-semigroup is a po-semigroup S containing an idempotent e such that \(x\leq e\) for all \(x\in S\) and such that if a,b,t\(\in S\) and \(a\leq b\), then ta\(\leq tb\) and at\(\leq bt\). A poe-semigroup S is said to be weakly commutative if for all x,y\(\in S\), there exists a positive integer n such that \((xy)^ n\leq yex\). S is called an Archimedean semigroup if for any a,b\(\in S\), there exists a positive integer m such that \(a^ m\leq xby\) for some x,y\(\in S\). In his paper ``On weakly commutative poe- semigroups'' [Semigroup Forum 34, 367-370 (1987; Zbl 0613.06009)], \textit{Niovi Kehayopulu} raised the following problem: Is each weakly commutative poe-semigroup S (uniquely) a semilattice of Archimedean subsemigroups? In the present paper it is shown that \(\mu =\{(a,b)\in S\times S:\) there exist positive integers m, n such that \(a^ m\leq xby\), \(b^ n\leq x_ 1by_ 1\) for some \(x,y,x_ 1,y_ 1\in S\}\) is a semilattice congruence such that each \(\mu\)-class is an Archimedean subsemigroup of S.