Weighting by iteration: iterations of \(n\) variables means based on subdivisions of the standard \((n-1)\)-simplex (Q2047332)

The author parametrizes means with the \((n-1)\) simplex \(\Delta_{n-1}=\{(w_1,w_2,\ldots,w_n)\in\mathbb{R}^n :w_i\geq 0, \sum_1^n w_i=1\}\). After an introduction of some concepts from simplicial homology and iterated systems of functions (termed ISF), the author considers lower means and studies their properties. Following this comes the barycentric iteration of means via a two step algorithm by first making an assignment on the face of each \(\Delta_{n-1}\) and iteratively applying the mean from previous assignments via \(M^{\frac{v_1+\cdots+v_n}{n}}(x)=M(M^{v_1}(x),\ldots,M^{v_n}(x))\) and studies the properties of these iterated means. It is shown that for symmetric means the orientation of the simplex does not matter. Next, weighting procedures in the iteration as well as weighting based on other subdivision in the manner of \textit{H. Freudenthal} [Ann. Math. (2) 43, 580--582 (1942; Zbl 0060.40701)] and properties of extension of the means are studied. The results are compared with those studied previously by other authors, in particular it is shown when the new methods coincide with \textit{J. Aczél}'s [Bull. Am. Math. Soc. 54, 392--400 (1948; Zbl 0030.02702)] weighting procedure for means and turn out to be continuous and scale invariant.