Recursive inset entropies of multiplicative type on open domains (Q1112266)

The authors investigate inset information measures, that is information measures depending both upon the events and the corresponding probabilities. More specifically, let B be a ring of subsets of a given set \(\Omega =\cup B\); moreover, let \(\Omega^ 0_ n=\{X=(x_ 1,...,x_ n)|\) \(x_ i\in B\setminus \{\emptyset \}\), \(x_ i\cap x_ j=\emptyset\) for all \(i\neq j\}\) \((x_ 1,...,x_ n\) are the ``events''), and \(\Gamma^ 0_ n=\{P=(p_ 1,...,p_ n)| \quad \sum^{n}_{i=1}p_ i=1,\) \(p_ i>0\) for all \(i\}\). For a fixed k, the authors present the general solution of all semisymmetric k-dimensional inset measures \(I_ n: \Omega_ n\times (\Gamma_ n)\to {\mathfrak R}\), which are recursive of multiplicative type \(\mu\). This result is of importance and interest since it contains as special cases nearly all recent characterization tolerances for generalized recursive information measures. If for example \(k=1\), then \[ I_ n(X;P)=f(x_ 1)\mu (p_ 1)+\sum^{n}_{i=2}g(x_ i)\mu (p_ i)=f(\cup^{n}_{i=1}x_ i)+\alpha (X;P) \] where \(\alpha (X;P)=\ell (p_ i)+\sum^{n}_{i=1}L(x_ i;p_ i)\) for \(\mu =1\), \(\alpha (X;P)=\sum^{n}_{i=1}\mu (p_ i)\ell (p_ i)\) for \(\mu\) additive and \(\alpha (X;P)=0\) otherwise. Here \(\ell\) is logarithmic, L is additive in the events and logarithmic in the probabilities, and f and g are arbitrary functions. The proof relies heavily (1) on the generalized fundamental equation of information of multiplicative type, which is another recent result of the authors [ibid. 32, 19-31 (1987; Zbl 0612.39006)]. (2) On the general solution of the non-probabilistic equation \(F(x\cup y,z)+F(x,y)=F(x\cup z,y)+F(x,z)\) for \((x,y,z)\in \Omega^ 0_ 3\). (To solve this equation, new techniques were developed.) For a survey of information measures satisfying weakly symmetric and generalized recursive properties we refer to the following paper of the reviewer [The fundamental equation of information and its generalizations, ibid. 33, 150-182 (1987)].