A representation theorem and a Lyapunov theorem for \(T_S\)-measures: The solution of two problems of Butnariu and Klement (Q1977813)

The paper deals with the Frank systems \(T_s\) \((s\in [0,\infty])\) of t-norms, i.e., with t-norms \(T\) satisfying the functional \(T(x,y)+ S(x,y)= x+ y\) \((x,y\in [0,1])\), where \(S\) is the dual conorm \((S(x,y)= 1- T(1- x,1-y))\). A special role plays here the Łukasiewicz t-norm \(T_\infty\) given by \(T_\infty(x,y)= \max(x+ y-1)\). \(T_s\)-measure is defined to be a function \(\mu:{\mathcal T}\to G\) defined on a \(T_s\)-clan \({\mathcal T}\) (i.e., \(0\in{\mathcal T}\), \(f\in {\mathcal T}\Rightarrow 1- f\in{\mathcal T}\), and \(f,g\in{\mathcal T}\Rightarrow fT_sg\in{\mathcal T}\)) with values in a normed commutative group such that \(\mu(0)= 0\), \(\mu(fTg)+ \mu(fSg)= \mu(f)+ \mu(g)\) and \(\mu\) is continuous from left. The representation theorem stated and proved in the paper states that every \(T_s\)-measure is the sum of a support measure \(\nu\) (i.e., \(\nu(f)= \varrho(\text{supp }f)\), where \(\varrho\) is a \(\sigma\)-additive measure) and a \(T_\infty\)-measure. The theorem presents a positive answer to a problem of Butnariu and Klement [see the book of \textit{D. Butnariu} and \textit{E. P. Klement}: ``Triangular norm-based measures and games with fuzzy coalitions'' (1993; Zbl 0804.90145)], solved partially by \textit{R. Mesiar} and \textit{M. Navara} [J. Math. Anal. Appl. 201, No. 1, 91-102 (1996; Zbl 0852.28009)]. Moreover, the theorem is used for the proof of the Lyapunov theorem: The set \(\mu({\mathcal T})\) is convex, if \(\mu:{\mathcal T}\to \mathbb{R}^n\) is a \(T_s\)-measure defined on a \(T_s\)-tribe \({\mathcal T}\) (i.e., \({\mathcal T}\) is \(T_s\)-clan closed with respect to infinite operations) that is \(\mu\)-chained, i.e., for every \(\varepsilon> 0\) and every \(f,g\in{\mathcal T}\) with \(f\leq g\) there exist \(f_0\leq f_1\leq\cdots\leq f_n\) in \({\mathcal T}\) such that \(f= f_0\), \(f_n= g\) and \(d_\mu(f_{i-1}, f_i)\leq \varepsilon\). Here \(d_\mu(f, g)= \sup\{|\mu(h)- \mu(k)|; f\wedge g\leq h\leq f\vee g\}\).