Fuzzy \(\varepsilon\)-subgroups. (Q2638198): Difference between revisions

Let \((S,\cdot)\) be a semigroup, \(G\) a subgroup of \(S\), \(L\) a complete lattice and \(\varepsilon\in L\). A function \(\mu\colon S\to L\) is a fuzzy \((G,\varepsilon)\)-subgroup (or simply an \(\varepsilon\)-subgroup) of \(S\) if the following conditions hold: (1) \(\forall x,y\in S\), \(\mu(x\cdot y)\geq\mu(x)\wedge\mu(y)\), where \(\wedge\) denotes minimum. (2) \(\forall x\in G\), \(\forall y\in S\setminus G\), \(\mu(x)\geq\varepsilon\geq\mu(y)\). (3) \(\forall x\in G\), \(\mu(x^{-1})\geq\mu(x)\). Using a special type of fuzzy partition, the authors prove that a fuzzy subsemigroup can be partitioned into a family of fuzzy \(\varepsilon\)-subgroups if and only if it is fuzzy completely regular. The authors show that if \(L\) is a completely distributive lattice and \(\mu\) is a fuzzy completely regular subsemigroup of \(S\), then there exists \(\varepsilon\in L\) and a family of fuzzy \(\varepsilon\)-subgroups of \(\mu\), \(\mathcal F=\{\mu_i\mid i\in L\}\), such that the following properties hold: (1) If \(i\neq j\), then \(\mu_i\) and \(\mu_j\) are disjoint. (2) \(\bigcup_{i\in I}\mu_i= \mu\). (3) \(\mu_i\circ \mu_j\subseteq\mu\) for all \(i,j\in L\). (4) \(\mu_i\circ\mu_i=\mu_i\) for all \(i\in L\). (5) \(\forall i,j\in L\), \(\exists k\in I\) such that \(\mu_i\circ\mu_j\subseteq\mu_k\) and \(\mu_j\circ\mu_i\subseteq\mu_k\).