A synthesis of multidimensional judgements consistent with linear transformations

If \(x_ 1,x_ 2,...,x_ m\in {\mathbb{R}}^ n\), the usual way of synthesizing \(x_ 1,x_ 2,...,x_ m\) into a single vector x is to consider a weighted quasiarithmetic mean of the form \[ M(x_ 1,x_ 2,...,x_ m)=\psi^{-1}[\sum^{m}_{i=1}P_ i\psi (x_ i)], \] where \(\psi\) : \({\mathbb{R}}^ n\to {\mathbb{R}}^ n\) is a continuous bijection and the weights \(P_ 1,...,P_ m\) are in (0,1), with \(\sum^{m}_{i=1}P_ i=1.\) The author determines which functions of the form above satisfy the condition \(M(A\cdot x_ 1,...,A\cdot x_ m)=A\cdot M(x_ 1,x_ 2,...,x_ m),\) for all \(x_ 1,x_ 2,...,x_ m\) in \({\mathbb{R}}^ n\) and all nonsingular \(n\times n\) matrices; namely he proves that M is the weighted arithmetic mean \(M(x_ 1,x_ 2,...,x_ m)=\sum^{m}_{i=1}P_ ix_ i\).

Recommendations

Cites work

Cited in

(1)

Synthesizing judgements: a functional equations approach

This page was built for publication: A synthesis of multidimensional judgements consistent with linear transformations

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2276697)