Relationship between ordered semihypergroups and ordered semigroups by using pseudoorder. (Q477719)

Most of the results on semigroups (ordered semigroups) can be transferred to hypersemigroups (ordered hypersemigroups). We never work directly on an hypersemigroup. If we want to obtain a result for an hypersemigroup (ordered hypersemigroup), then we have to solve it first for a semigroup (ordered semigroup) and transfer its proof to hypersemigroup (ordered hypersemigroup). On the other hand, if we prove a problem for a semigroup (ordered semigroup) we normally publish it and do not keep it to publish it for an hypersemigroup (ordered hypersemigroup). So every published paper on an hypersemigroup (ordered hypersemigroup) comes from a published paper on a semigroup (ordered semigroup); that's why, in every paper on hypersemigroups (ordered hypersemigroups) a proper reference list is needed. The paper under review is based on the papers by \textit{N. Kehayopulu} and \textit{M. Tsingelis} [Semigroup Forum 50, No. 2, 161--177 (1995; Zbl 0823.06010)] and [Semigroup Forum 50, No. 3, 389--392 (1995; Zbl 0828.06010)]; mainly on the first one. It is well known that if \(S\) is a semigroup and \(\sigma\) a congruence on \(S\), then the set \(S/\sigma:=\{(a)_\sigma\mid a\in S\}\) of all \(\sigma\)-classes on \(S\) is a semigroup. In the paper ``On subdirectly irreducible ordered semigroups'' mentioned above [Zbl 0823.06010], they put the following problem: If \((S,\cdot,\leq)\) is an ordered semigroup and \(\sigma\) a congruence on \(S\), then is the set \(S/\sigma\) an ordered semigroup? First they gave an example that the probably order \(\preceq\) on \(S/\sigma\) defined by \[ \begin{aligned} \preceq: & =\{(t,z)\in S/\sigma\times S/\sigma \mid \exists\;(a,b)\in\;\leq \text{ such that } t=(a)_{\sigma}, z=(b)_{\sigma}\}\\ & =\{((x)_{\sigma},(y)_{\sigma}) \mid \exists\; a\in (x)_{\sigma}, b\in (y)_{\sigma} \text{ such that } (a,b)\in\;\leq\}\end{aligned} \] is not an order in general. To give an answer to this problem, they introduced the concept of \textit{pseudoorder} and they proved that if \((S,\cdot,\leq)\) is an ordered semigroup and \(\sigma\) a pseudoorder on \(S\), then the relation \(\overline \sigma \) on \(S\) defined by \[ \overline\sigma:=\{(a,b)\in S\times S \mid (a,b)\in\sigma \text{ and } (b,a)\in\sigma\} \] is a congruence on \(S\) and \(S/\overline\sigma\) is an ordered semigroup. In the paper under review the authors transferred the results and the problem mentioned above from ordered semigroups to ordered hypersemigroups. And as to a congruence of a semigroup correspond two concepts of congruences of an hypersemigroup (the authors call one of them ``regular relation'' and the other ``strongly regular relation''), they dealt with one of them (the strongly regular relation) (Theorem 4.2) and put the other case (the case of regular relation) as an open problem. In Theorem 2.1 condition (2) has no sense; the same for condition (2) in Theorem 2.2. They use the Theorem 2.2 throughout the paper but this theorem should be corrected. The examples on ordered hypersemigroups given by a table of multiplication and an order also come from ordered semigroups. Even the expression ``we give the covering relation and the figure of \(S\)'' does not belong to the authors and a reference should be certainly given.