Orador :: Prof Charles Johnson (The College of William and Mary, VA, USA)
Título :: Reciprocal Matrices and Ranking
Data :: 3 de Outubro de 2023, 12h
Local ::
presencial: Sala de Seminários do DMAT
online: https://videoconf-colibri.zoom.us/j/93344876599?pwd=ejkrZ3FpYkh0NE1YV3NCdGFMOGw3QT09
Resumo:: If $A=(a_{ij})$ is a positive matrix, it is called "reciprocal" if $a_{ji}=\frac{1}{a_{ij}}$, all $1\leq i,j,\leq n$. The entries are intended as pairwise, ratio comparisons among $n$ items. We mention several examples. The goal is to deduce a positive vector $w$ from $A$ to be used either ordinally or cardinally to rank the $n$ items. Such vector should be chosen from among "efficient" vectors, which will be defined. The efficient vectors will be discussed, as well as applications.